A comprehensive, step-by-step guide to learning Machine Learning from absolute basics to advanced topics
Structured path for beginners • 26 Learning Modules • 23 Real-World Projects • Deployment and MLOps topics
Start Here • How to Use the Study Hub • Foundation & job readiness • Get Started • Learning Path • Projects • System Design • Full-stack AI • Contribute
Watch the full playlist on YouTube
- Zero to Advanced Path: Complete learning journey from basics to advanced ML
- Beginner-Friendly: No prior experience needed. We cover everything
- Hands-On Learning: 23 practical projects to build your portfolio
- Production Topics: Deployment, MLOps, and real-world failure modes (not a guarantee that one pass makes you production-ready)
- Well-Organized: Logical progression with clear learning objectives
- Community-Driven: Open source, contributions welcome!
Built for: Students, career switchers, and self-learners who want a structured ML path. Outcomes depend on practice, projects, and the job market. This repo does not guarantee a hire.
Choose your path! This repository prepares you for multiple ML/AI careers. Select your target role to see a customized learning path:
| Role | Focus | Est. Time | Key Modules | Full Guide |
|---|---|---|---|---|
| Data Analyst | Insights & Reports | 8-12 months | 00, 01, 19, 20, 21 | View Guide → |
| Data Scientist | Predictive Models | 13-20 months | 00-08, 15, 19-21 | View Guide → |
| ML Engineer | Production ML | 17-26 months | 00-10, 13-14, 19-21 | View Guide → |
| LLM Engineer | Language Models | 17-24 months | 00-01, 05, 09-10, 12, 25, 13-14, 19 | View Guide → |
| GenAI Solution Architect | Production GenAI | 15-21 months | 00-01, 02, 05, 09-10, 12, 25, 13-14, 19 | View Guide → |
| Computer Vision Engineer | Image Processing | 16-25 months | 00-01, 04-05, 09-11, 13-14, 19, 21 | View Guide → |
| AI Engineer | Generalist AI | 25-38 months | 00-15, 19-21, 22-24, 25 | View Guide → |
| Data Engineer | Data Infrastructure | 14-20 months | 00-01, 13-14, 19-20 | View Guide → |
| MLOps Engineer | ML Operations | 16-24 months | 00-01, 05, 09-10, 12, 25, 13-14, 19 | View Guide → |
| Research Scientist | Novel methods, careful experiments, clear write-ups | 24-34 months | 00-12, 15, 19, 21, 22-24, 25 | View Guide → |
| BI Analyst | Business Intelligence | 10-15 months | 00-01, 19-21 | View Guide → |
| Full-Stack AI Engineer | End-to-end AI products | 12-24 months (15-25 hrs/week, parallel with blueprint) | 00-01, 19, 25 + In-repo lessons + Blueprint A–H | View Guide → |
Foundational side track: System Design for Beginners: 31 lessons (22 foundations + 9 interview-style designs) covering HTTP, TCP, DNS, caching, load balancing, CAP, sharding, and message queues, including a capstone-style design sequence (rate limiter, social feed, chat, video, file storage, maps, KV store, message queue, and a real-time ML feature store). Useful for ML Engineer, MLOps, Data Engineer, AI Engineer, and Full-Stack AI Engineer paths. Pairs with the ML System Design Guide, which applies these foundations to ML services. For application backend engineering (auth, Postgres, queues, ops), use the Backend Engineer Roadmap alongside Phases B/C/E of the Full-Stack Track.
Time assumptions: Ranges in the Est. Time column assume about 10–15 hours/week for a steady part-time pace and 30–40 hours/week for an intensive full-time pace, unless a row states otherwise (for example, Full-Stack AI Engineer calls out parallel blueprint hours). Time-to-role also depends on prior software experience, portfolio quality, and hiring markets. Finishing modules is not the same as getting hired.
SQL timing: Module 19 is Stage 1.5 (parallel with Modules 01–02) for most roles, not after GenAI. Module lists like 19-21 mean you need all three modules, not that SQL comes last.
What each path includes:
- Career Roadmap Guide roles: module priorities, suggested projects, essential resources, skills checklists, and time guidance grounded in this repo’s 00–25 folders.
- Full-Stack AI Engineer: same style of overview inside the Career Roadmap Guide, plus in-repo lesson chapters and the companion Full-Stack AI Engineer Blueprint for TypeScript through deployment and AI integration. Most hands-on engineering work is outside the numbered module folders; use the lessons + blueprint phases and free links for day-to-day study.
The full roadmap guide follows the same module numbering and Stage 6 vs. module 15 (time series) branch as the Learning Path Overview above.
View Complete Career Roadmap Guide →
If you want ML plus product engineering (TypeScript, APIs, databases, Next.js, containers, and AI features in production), use this sequence:
- Skim the role overview: Full-Stack AI Engineer
- Work through in-repo lesson chapters (Phases A–H)
- Follow the Full-Stack AI Engineer Blueprint for deliverables, external free links, and portfolio milestones
Keep progressing through the numbered 00–25 modules for core ML depth in parallel when you can.
If you're heading toward ML Engineer, MLOps, Data Engineer, AI Engineer, or Full-Stack AI Engineer, you'll need backend system design vocabulary (HTTP, TCP, caching, load balancing, CAP, sharding, message queues). Read these in order:
- System Design for Beginners: 22 foundation lessons + 9 interview-style designs (~8 hours total).
- ML System Design Guide: how those fundamentals apply to ML services (latency, serving, drift, MLOps).
- Interview Preparation Guide: practice questions in the System Design section.
You don't need to finish all 26 ML modules first. The system design track is parallel and pairs well with module 13 (deployment), 14 (MLOps), and any project phase (16–18).
- Overview
- Learning Path
- Prerequisites
- Getting Started
- Repository Structure
- Resources
- YouTube Playlist
- Career Paths
- Full-stack AI quick start
- System design foundations
- Contributing
This repository provides a structured learning path for machine learning, organized in a logical progression from fundamentals to advanced topics. Each module includes:
- Clear explanations of concepts
- Hands-on exercises with solutions
- Practical projects to reinforce learning
- Additional resources for deeper understanding
Follow stages, not folder numbers. Folders
13–14(deployment) appear before19(SQL) on disk. That is intentional for repo layout, not teaching order. Use START-HERE.md and FOUNDATION_AND_JOB_READINESS.md. SQL is Stage 1.5 for most job tracks.
Note on numbering: Folder names use Module 00–25 (for example 09-neural-networks-basics). Stages in the table below are the recommended learning sequence. A stage can span several modules. They are not the same label. Some modules can be learned in parallel depending on your goals. See each module README for prerequisites.
Time Estimates: Realistic completion time is 15–22 months full-time (30–40 hrs/week) or 30–39 months part-time (10–15 hrs/week) for full coverage of all 26 modules and 23 projects. See the FAQ section for a stage-by-stage breakdown.
Structured courses usually stack Python, then math and stats, then tabular ML, then evaluation and feature work, then deep learning and GenAI, then SQL and storytelling, then production and MLOps, then electives. This repository follows the same logic but groups topics by module folder instead of week numbers. Stage 0 covers programming plus calculus, linear algebra, and stats. Stage 1 covers NumPy, Pandas, visualization, EDA, APIs, and SQL access. Stages 2–4 cover classical supervised and unsupervised ML. Stages 5–7 cover neural nets, specialized deep learning, and LLM-era tooling. Modules 13–14 cover deployment and experiment discipline. Module 15 and modules 22–24 are branch tracks for time series, RL, graphs, and audio. If your external syllabus mentions MLE and odds for logistic models, ROC vs PR, data leakage, Airflow or Kubernetes monitoring, or Kaggle-style iteration, look first in modules 04, 05, 06–07, 14, and the project folders.
| Stage | Modules | Focus Area | Est. Time (Full-Time) | Est. Time (Part-Time) |
|---|---|---|---|---|
| Stage 0 | 00 | Foundation (Python, Math) | 2-3 months | 4-6 months |
| Stage 1 | 01 | Data Fundamentals | 2-3 months | 4-6 months |
| Stage 1.5 | 19 | SQL & Databases (parallel with 1–2 for job tracks) | 1-2 months | 2-3 months |
| Stage 2 | 02-05 | ML Basics | 2-3 months | 4-6 months |
| Stage 3 | 06-07 | Advanced ML | 1-2 months | 2-4 months |
| Stage 4 | 08 | Unsupervised Learning | 1 month | 2 months |
| Stage 5 | 09-10 | Deep Learning Fundamentals | 2 months | 4 months |
| Stage 6 | 11-12 | Vision & language (specialized DL) | 3-4 months | 6-8 months |
| Branch | 15 | Time series (pick one path) | 0.5-1 month | 1-2 months |
| Stage 7 | 25 | Generative AI & Modern LLMs | 1-2 months | 2-4 months |
| Stage 7.5 | 20-21 | Imbalanced Data, Explainability | 1-2 months | 2-4 months |
| Stage 8 | 13-14 | Production & MLOps | 2-3 months | 4-6 months |
| Stage 9 | 16-18 | Projects (23 total) | 4-6 months | 8-12 months |
| Stage 10 | 22-24 | Advanced Specialized Topics (RL, GNNs, Audio) | 2-3 months | 4-6 months |
| Total | 26 modules | Complete Path | 15-22 months | 30-39 months |
Goal: Build the mathematical and programming foundation needed for ML
- 00-prerequisites
- AI programming with Python
- Variables, operators, control flow, strings (indexing/slicing), lists, tuples, sets, and dictionaries;
importand modules;break/continue/pass - Functions and functional programming;
*args/**kwargs; nested functions and scope - Introduction to OOP: classes and objects; methods; magic methods; inheritance, polymorphism, encapsulation, abstraction
- Iterators and generators; lambda functions;
map,filter, andreduce - File handling: text and binary I/O; read position with
tell/seek; context managers (with); JSON and pickle (when each is appropriate) - Decorators; namespaces and the LEGB scope rule
- Practice problems based on file handling
- Exceptions:
try/except/else/finally; specific errors,raise, and custom exception types - Project: Movie script generator
- Time Complexity & Algorithm Efficiency (Big O notation)
- GUI Development with tkinter
- Essential Mathematics (Linear Algebra including Tensors, Statistics, Calculus basics)
- Computational Math: Python code examples alongside theory for better intuition
- Environment Setup (Python, Jupyter, Virtual Environments, Git & GitHub basics)
Goal: Master data manipulation and visualization
Core path (finish before Stage 2):
- 01-python-for-data-science
- NumPy fundamentals, array operations, broadcasting, linear algebra basics
- Pandas Series and DataFrame, indexing, GroupBy, merge/join, missing data
- Matplotlib and Seaborn for core plots
- Exploratory data analysis and data wrangling
Optional depth (add when a job or project needs them):
-
Polars and Dask for larger datasets
-
Plotly, Dash, Streamlit, Flask, and Tableau for apps and dashboards
-
ETL with AWS RDS, Selenium scraping, regular expressions
-
Project: Build a first model end-to-end (see
02-introduction-to-ml/first-ml-project-tutorial.mdand beginner projects in module 16)
Goal: Query and integrate relational data, required for most analyst and data scientist roles
- 19-sql-database-fundamentals
- Database fundamentals, DDL/DML, joins, subqueries, window functions
- CTEs, data cleaning in SQL, Python integration (SQLAlchemy/pandas)
- NoSQL overview (MongoDB, Redis, etc.)
- Suggested timing: Start after Module 01 lessons 2–3; do not wait until after GenAI (Module 25)
Goal: Understand core ML concepts and algorithms
Machine learning for beginners (maps with runnable examples: Introduction → Regression → Classification; EDA and preprocessing in EDA and Feature engineering; practice in Beginner projects)
-
02-introduction-to-ml
- Machine learning for beginners framing; data types in ML; population vs sample
- Descriptive statistics fundamentals: mean, median, mode, variance, standard deviation (with code)
- What is Machine Learning? Types (Supervised, Unsupervised, Reinforcement); ML workflow and best practices; vocabulary for batch vs online updates and instance-based vs model-based learners (see guide)
- Projects: first end-to-end model (first ML project)
-
03-supervised-learning-regression
- Linear regression (simple and multiple) and evaluation metrics (MSE, RMSE, MAE, R²)
- Polynomial and multiple regression techniques; regularization (Ridge, Lasso)
- Optimization lens: gradient descent intuition and batch vs stochastic vs mini-batch updates (ties to Stage 0 calculus and Stage 5 deep learning)
- Data distribution analysis: percentiles, histogram, boxplot (see guide)
- Feature relationships: scatter plots, covariance, correlation (with regression workflow)
- End-to-end regression project path and saving a model (train → evaluate → persist); full walkthrough: regression project tutorial; deployment: model deployment module
- Statistical regression analysis (statsmodels: TSS, RSS, ESS, F-statistic, p-values, confidence intervals)
-
04-supervised-learning-classification
- Logistic regression (including multiclass / softmax framing), maximum likelihood view of loss, and K-Nearest Neighbours (distance-based learning); Naive Bayes (probabilistic classification)
- Decision Trees, Random Forests, SVM (multiclass strategies; Accuracy, Precision, Recall, F1, ROC-AUC); advanced track in same guide: entropy / information gain / rules, SVM margin–hyperplane–kernels, tree depth vs overfitting, tree vs forest, supervised model comparison (map)
- Exploratory data analysis and data cleaning (tie-in with Stage 1 EDA); feature relationship analysis (scatter, covariance, correlation)
- Bias auditing and fairness (Fairlearn, demographic parity, equalized odds)
-
05-model-evaluation-optimization
- Train/Validation/Test Split
- Cross-Validation (K-Fold, Stratified, Leave-One-Out, Time Series Split)
- ROC and PR curves, thresholds, and when PR beats ROC on imbalanced problems
- Data leakage: how it creeps in and how to design splits that prevent it
- Hyperparameter Tuning (Grid Search, Random Search, Bayesian Optimization with Optuna)
- Bias-Variance Tradeoff
- Learning Curves
- Overfitting and Underfitting
- Model Calibration (Probability Calibration, Platt Scaling, Isotonic Regression)
Goal: Explore ensemble methods and advanced techniques
Advanced machine learning (trees, SVM theory, supervised comparison: Classification guide; ensembles: Ensemble methods; clustering: Unsupervised)
-
06-ensemble-methods
- Random Forest and bagging; boosting (AdaBoost, Gradient Boosting, XGBoost, LightGBM, CatBoost); stacking; voting
- Map: Advanced ML curriculum
-
07-feature-engineering
- Feature Selection
- Feature Transformation
- Handling Categorical Variables (One-Hot, Label, Target, Frequency, WOE Encoding)
- Feature Scaling and Normalization
- Dimensionality Reduction (PCA)
- Advanced Discretization (Decision Tree-based binning, custom strategies)
- Decision Tree Visualization with dtreeviz
- Comprehensive sklearn Pipeline and ColumnTransformer Guide
Goal: Make sense of data when nobody handed you labels.
Module 08 is where you work without a target column. You cluster, reduce dimensions, spot outliers, and explain why a grouping is plausible. These skills carry into EDA, modeling, and research writing. Follow the unsupervised walkthrough when you want structure beyond trying k=3. Unsupervised learning map.
- 08-unsupervised-learning. K-Means, hierarchical and density-based clustering, validation and plots, PCA / t-SNE / SVD, LDA (supervised dimensionality reduction) beside PCA, anomaly detection, association rules (e.g. Apriori / market-basket style).
Goal: Build intuition that survives new architectures.
Most breakthroughs are still “layers + nonlinearity + data + optimization.” Here you learn to read training curves, sanity-check gradients, and see why a linear model fails before frameworks hide the details. If you aim toward a research-scientist path, treat this stage as lab notebook habits. Run small experiments, change one thing at a time, and keep runs reproducible.
-
09-neural-networks-basics. From linear limits to MLPs: activations, losses, backpropagation as chain rule, optimizers (SGD family through Adam), regularization, initialization, and what actually breaks training. Study map →
-
10-deep-learning-frameworks. Turn math into working code: PyTorch tensors and autograd, real training loops, Keras ergonomics, saving models, and shipping-friendly formats when you need them. Study map → · Practice builds → (MNIST, churn, tabular regression)
Goal, see the world as tensors. Pixels and tokens.
Vision and language split into specialist toolkits here, but the mindset stays the same. Use the right inductive bias (convolution for locality, attention for pairing), build careful data pipelines, and evaluate honestly. Compare baselines, ablate one change at a time, and write down what surprised you.
Where to dive: Vision roadmap · Language roadmap · PyTorch bridge for CV + NLP · Open-ended projects
-
11-computer-vision. Convolutions, CNN families, augmentation, transfer learning, detection, segmentation, and ViT-style ideas.
-
12-natural-language-processing. RNN to LSTM/GRU, attention, transformers, Hugging Face workflows, retrieval, and sentiment on messy text.
Goal: Work with foundation models, not against them.
You rarely train a frontier LLM from scratch. You compose prompts, retrieval, tools, evaluation harnesses, and guardrails. This stage is about shipping ideas quickly while keeping rigor. Trace failures, measure hallucinations, version prompts like code, and know when a vector DB is the right memory.
- 25-generative-ai-llms. Prompt patterns, embeddings, vector search, RAG pipelines, agents, and production patterns. Deeper reads in GenAI overview, RAG guide, LangChain, Agents.
Pacing: Roughly 1–2 months full-time (2–4 part-time), often best after Stage 6 or alongside NLP if you learn by building products early.
Forecasting and monitoring problems break the usual “shuffle and split” story. Treat 15 as a side quest you can start after you are comfortable with regression or after RNNs: many research and product teams touch time series even if it is not their main title.
- 15-time-series-analysis. Trend and seasonality, classical models (ARIMA family), deep sequence models, feature design for temporal data, and evaluation that respects time ordering. Numbering note: module 15 sits after 13–14 in the repo but fits logically here or alongside Stages 5–6.
Goal: Master critical skills for real-world ML applications
Note: These modules can be learned in parallel with other phases or integrated earlier in your learning journey:
-
DSA (recommended for interviews): Follow the DSA Course (Python) alongside these modules to build strong coding + problem-solving fundamentals
-
SQL is Stage 1.5 (Module 19), not part of this stage; see above
-
Imbalanced Data is most useful after Stage 2 (Classification) when you encounter real-world datasets
-
Model Explainability is valuable after Stage 3 (Advanced ML) when working with complex models
-
20-handling-imbalanced-data
- Understanding Imbalanced Data Problems
- Resampling Techniques (SMOTE, ADASYN, Undersampling, Combined)
- Algorithm-Level Solutions (Class Weights, Threshold Tuning, Cost-Sensitive Learning)
- Appropriate Evaluation Metrics (PR-AUC, F1-Score)
- Complete Workflow Examples
- Best Practices and Common Pitfalls
- Suggested Timing: Best learned after Stage 2 (Classification) or Stage 3 (Advanced ML)
-
21-model-explainability
- Feature Importance Methods (Tree-based, Permutation)
- SHAP (SHapley Additive exPlanations): Tree, Kernel, and Advanced
- LIME (Local Interpretable Model-agnostic Explanations)
- Partial Dependence Plots (PDP) and ICE Plots
- Model Interpretation Best Practices
- Regulatory Compliance and Ethical AI
- Suggested Timing: Best learned after Stage 3 (Advanced ML) or Stage 5 (Deep Learning)
Goal: Learn to deploy ML models in production
Note: Modules 13-14 are numbered before modules 15-18 but logically come after specialized deep learning and generative AI topics. Learn these after you have built and trained models.
-
13-model-deployment
- Model Serialization
- REST APIs with Flask/FastAPI
- Docker for ML
- Cloud Deployment (AWS, GCP, Azure, Hugging Face Spaces)
- Hugging Face Spaces: Free model hosting for portfolio projects
- Production Server Setup (NGINX, SSL/TLS, Domain Configuration)
- Security Best Practices (Rate Limiting, Authentication, Input Validation)
- AWS SageMaker Comprehensive Guide
- A/B Testing (Statistical Significance, Multi-Armed Bandits, Sequential Testing)
- Model Monitoring
-
14-mlops-basics
- Version Control for ML (DVC, MLflow)
- CI/CD for ML (e.g. GitHub Actions) and reproducible project templates (Cookiecutter)
- Experiment Tracking (MLflow, Weights & Biases with advanced features)
- Model Registry
- Weights & Biases: Hyperparameter sweeps, model artifacts, team collaboration
- Orchestration & scale (topics and links in module guides): DAG-style pipelines (e.g. Airflow, Kubeflow), containers and Kubernetes basics, rollout patterns (canary / blue–green), metrics and alerting (e.g. Prometheus / Grafana) where you operationalize models
- Apache Kafka for Data Streaming
- Apache Spark for Big Data Processing
Goal: Apply knowledge through real-world projects
Note: Projects live in modules 16–18. Start beginner projects after Stage 2, intermediate after Stages 3–4, and advanced after Stages 7–8.
Delivery: Beginner projects include runnable code in the repo. Intermediate and advanced projects ship as detailed README briefs. You build notebooks and scripts yourself.
| Category | Count | Prerequisites | Est. Time | Delivery |
|---|---|---|---|---|
| Beginner | 6 | Stages 0–2 | 2–3 weeks | Code included |
| Intermediate | 8 | Stages 0–4 | 4–6 weeks | README brief |
| Advanced | 9 | Stages 0–7 | 8–12 weeks | README brief |
| Total | 23 | , | 4–6 months | Mixed |
| # | Project Name | Skills | Time | Status |
|---|---|---|---|---|
| 1 | House Price Prediction | Regression, Feature Engineering, EDA | 2-3 days | ✓ Available |
| 2 | Iris Flower Classification | Classification, EDA, Multiple Algorithms | 1 day | ✓ Available |
| 3 | Titanic Survival Prediction | Classification, Data Cleaning, Feature Engineering | 2-3 days | ✓ Available |
| 4 | Spam Email Detection | Text Classification, NLP Basics | 2-3 days | ✓ Available |
| 5 | Wine Quality Prediction | Regression, Feature Engineering | 2-3 days | ✓ Available |
| 6 | Customer Data Dashboard with Streamlit | Data Visualization, Streamlit | 3-5 days | ✓ Available |
Prerequisites: Complete Stages 0-2 before starting
| # | Project Name | Skills | Time | Status |
|---|---|---|---|---|
| 1 | Handwritten Digit Recognition (MNIST) | Neural Networks, Image Processing | 3-5 days | ✓ Available |
| 2 | Customer Churn Prediction | Classification, Imbalanced Data, Business Metrics | 4-5 days | ✓ Available |
| 3 | Movie Recommendation System | Collaborative Filtering, Content-Based | 5-7 days | ✓ Available |
| 4 | Credit Card Fraud Detection | Anomaly Detection, Imbalanced Data | 4-5 days | ✓ Available |
| 5 | Customer Segmentation | Clustering, Unsupervised Learning | 3-4 days | ✓ Available |
| 6 | Time Series Forecasting | Time Series Analysis, ARIMA, LSTM | 5-7 days | ✓ Available |
| 7 | Feature Engineering Mastery | Feature Engineering, Advanced Techniques | 4-5 days | ✓ Available |
| 8 | Ensemble Methods Comparison | Ensemble Methods, Model Comparison | 3-4 days | ✓ Available |
Prerequisites: Complete Stages 0–4; modules 20–21 (Stage 7.5) help for churn, fraud, and ensemble projects
Delivery: Each project includes starter.py, see intermediate projects README.
| # | Project Name | Skills | Time | Status |
|---|---|---|---|---|
| 1 | Image Classification (CIFAR-10) | CNNs, Transfer Learning, Data Augmentation | 1-2 weeks | ✓ Available |
| 2 | Sentiment Analysis on Reviews | NLP, RNNs/LSTMs, Transformers | 1-2 weeks | ✓ Available |
| 3 | Time Series Forecasting (Advanced) | Advanced Time Series, Deep Learning | 1-2 weeks | ✓ Available |
| 4 | LLM Chatbot & RAG System | Modern LLMs, RAG, Vector Databases, LangChain | 2-3 weeks | ✓ Available |
| 5 | Object Detection | Computer Vision, YOLO, R-CNN | 1-2 weeks | ✓ Available |
| 6 | End-to-End ML Pipeline | Full ML Pipeline, MLOps | 2-3 weeks | ✓ Available |
| 7 | Generative Model (GAN/VAE) | GANs, VAEs, Image Generation | 1-2 weeks | ✓ Available |
| 8 | Model Explainability & Interpretability | SHAP, LIME, Model Interpretation | 1-2 weeks | ✓ Available |
| 9 | Model Deployment & Serving | Model Deployment, APIs, Cloud | 1-2 weeks | ✓ Available |
Optional Capstones (Portfolio-Grade, GDPR-Safe):
- Capstone Blueprints → (ML Engineer, LLM/RAG, Data/Analytics)
Prerequisites: Complete Stages 0-7 recommended for full benefit
Goal: Push into domains where structure is not a tidy table of rows.
Note: Reinforcement learning, graphs, and audio each need solid deep-learning fundamentals. Treat 22–24 as electives you pick when a problem pulls you there, after Stage 5–6 (and often after you have shipped at least one non-trivial model).
-
22-reinforcement-learning
- Markov Decision Processes (MDPs)
- Value-Based Methods (Q-Learning, DQN)
- Policy-Based Methods (REINFORCE, Policy Gradients)
- Actor-Critic Methods
- Deep Reinforcement Learning
- Multi-Agent RL, Hierarchical RL, Imitation Learning
- Applications: Game Playing, Robotics, Recommendation Systems
-
23-graph-neural-networks
- Graph Fundamentals and Representations
- Message Passing in GNNs
- Graph Convolutional Networks (GCNs)
- Graph Attention Networks (GATs)
- GraphSAGE and Other Architectures
- Applications: Social Networks, Recommendation Systems, Molecular Analysis
-
24-audio-speech-processing
- Audio Signal Fundamentals (Waveforms, Spectrograms, MFCCs)
- Speech Recognition (ASR): CTC, Attention-based, Whisper
- Text-to-Speech (TTS): Neural TTS, Voice Cloning
- Audio Classification (Music, Events, Emotions)
- Music Generation
- Voice Processing (VAD, Speaker ID, Enhancement)
Before starting, you should have:
- Basic computer literacy
- Willingness to learn and practice
- A computer with internet connection
Note: No prior programming or math experience required! We'll cover everything you need.
git clone https://github.com/NabidAlam/road-to-machine-learning.git
cd road-to-machine-learning# Install Anaconda from https://www.anaconda.com/products/individual
# Create a new environment
conda create -n ml-env python=3.10
conda activate ml-env
# Install required packages
pip install -r requirements.txt# Create virtual environment
python -m venv ml-env
# Activate (Windows)
ml-env\Scripts\activate
# Activate (Mac/Linux)
source ml-env/bin/activate
# Install required packages
pip install -r requirements.txtpip install jupyter notebook
jupyter notebookQuick Start Option: New here? Read START-HERE.md first. Want to see ML in action immediately? Try GETTING_STARTED.md for a 30-minute first project.
Full Learning Path: Follow the modules in order:
- Start with
00-prerequisites - Progress through each module sequentially
- Complete exercises and projects
- Practice, practice, practice!
Note: All learning modules now include comprehensive detailed guides with code examples, exercises, and solutions. Beginner projects are fully available with READMEs and code. Intermediate and advanced projects have detailed READMEs with instructions.
road-to-machine-learning/
00-prerequisites/
01-python-basics.md (includes time complexity, iterators/generators)
02-linear-algebra.md
03-statistics-probability.md
04-calculus.md
05-environment-setup.md
prerequisites-advanced-topics.md
prerequisites-project-tutorial.md
prerequisites-quick-reference.md
README.md
01-python-for-data-science/
01-numpy.md
02-pandas.md
03-visualization.md (includes Plotly & Dash)
04-exploratory-data-analysis.md
05-data-sources-and-integration.md (includes ETL with AWS RDS, Advanced Selenium)
06-regular-expressions-text-processing.md (includes advanced regex: lookahead, lookbehind, back references)
07-advanced-data-wrangling.md
08-working-with-dates-times.md
09-streamlit-dashboards.md
10-flask-web-development.md
11-tableau-visualization.md
python-for-data-science-advanced-topics.md
python-for-data-science-project-tutorial.md
python-for-data-science-quick-reference.md
README.md
02-introduction-to-ml/
introduction-to-ml.md
introduction-to-ml-advanced-topics.md
introduction-to-ml-quick-reference.md
ml-terminology.md
problem-identification-algorithm-selection.md
first-ml-project-tutorial.md
common-pitfalls-best-practices.md
README.md
03-supervised-learning-regression/
regression.md
regression-advanced-topics.md (includes statsmodels)
regression-project-tutorial.md
regression-quick-reference.md
README.md
04-supervised-learning-classification/
classification.md (includes dtreeviz visualization)
classification-advanced-topics.md
classification-project-tutorial.md
classification-quick-reference.md
README.md
05-model-evaluation-optimization/
evaluation-optimization.md
evaluation-optimization-advanced-topics.md
evaluation-optimization-project-tutorial.md
evaluation-optimization-quick-reference.md
README.md
06-ensemble-methods/
ensemble-methods.md
ensemble-methods-advanced-topics.md
ensemble-methods-project-tutorial.md
ensemble-methods-quick-reference.md
README.md
07-feature-engineering/
feature-engineering.md (includes WOE encoding, advanced discretization)
feature-engineering-advanced-topics.md (includes sklearn Deep Dive: Estimators, Mixins, Composite Transformers, FeatureUnion)
feature-engineering-project-tutorial.md
feature-engineering-quick-reference.md
README.md
08-unsupervised-learning/
unsupervised-learning.md
unsupervised-learning-advanced-topics.md (includes SVD - Singular Value Decomposition with PCA connection)
unsupervised-learning-project-tutorial.md
unsupervised-learning-quick-reference.md
README.md
09-neural-networks-basics/
neural-networks.md
neural-networks-advanced-topics.md
neural-networks-project-tutorial.md
neural-networks-quick-reference.md
README.md
10-deep-learning-frameworks/
deep-learning-frameworks.md
deep-learning-frameworks-advanced-topics.md
deep-learning-frameworks-project-tutorial.md
deep-learning-frameworks-quick-reference.md
README.md
11-computer-vision/
computer-vision.md
computer-vision-advanced-topics.md
computer-vision-project-tutorial.md
computer-vision-quick-reference.md
README.md
12-natural-language-processing/
nlp.md
nlp-advanced-topics.md
nlp-project-tutorial.md
nlp-quick-reference.md
README.md
13-model-deployment/
deployment.md (includes FastAPI advanced features: type checking, dependency injection, background tasks)
deployment-advanced-topics.md (includes AWS SageMaker comprehensive guide)
deployment-project-tutorial.md
deployment-quick-reference.md
README.md
14-mlops-basics/
mlops.md (includes Cookiecutter for Data Science)
mlops-advanced-topics.md (includes Apache Kafka, Apache Spark, Feature Stores)
mlops-project-tutorial.md
mlops-quick-reference.md
README.md
15-time-series-analysis/
time-series-analysis.md
time-series-advanced-topics.md
time-series-project-tutorial.md
time-series-quick-reference.md
resources.md
exercises/README.md
README.md
16-projects-beginner/
projects-beginner.md
projects-beginner-advanced-topics.md
projects-beginner-project-tutorial.md
projects-beginner-quick-reference.md
project-01-house-price-prediction/
project-02-iris-classification/
project-03-titanic-survival/
project-04-spam-detection/
project-05-wine-quality/
project-06-customer-dashboard/
(each: code + README)
README.md
17-projects-intermediate/
project-01-mnist-digit-recognition/
project-02-customer-churn/
project-03-movie-recommendation/
project-04-fraud-detection/
project-05-customer-segmentation/
project-06-time-series-forecasting/
project-07-feature-engineering/
project-08-ensemble-comparison/
(each: README + instructions)
README.md
18-projects-advanced/
capstones/ (README + capstone-ml-engineer.md, capstone-llm-rag-engineer.md, capstone-data-analytics-sql-ml.md)
project-01-cifar10-classification/
project-02-sentiment-analysis/
project-03-time-series-forecasting/
project-04-chatbot/
project-05-object-detection/
project-06-end-to-end-pipeline/
project-07-generative-model/
project-08-model-explainability/
project-09-model-deployment/
(each: README + instructions)
README.md
19-sql-database-fundamentals/
sql-database.md
sql-database-advanced-topics.md (includes NoSQL: MongoDB, Redis, Cassandra, Neo4j)
sql-database-project-tutorial.md
sql-database-quick-reference.md
README.md
20-handling-imbalanced-data/
imbalanced-data.md
imbalanced-data-advanced-topics.md
imbalanced-data-project-tutorial.md
imbalanced-data-quick-reference.md
README.md
21-model-explainability/
model-explainability.md
model-explainability-advanced-topics.md
model-explainability-project-tutorial.md
model-explainability-quick-reference.md
README.md
22-reinforcement-learning/
reinforcement-learning.md
reinforcement-learning-advanced-topics.md
reinforcement-learning-project-tutorial.md
reinforcement-learning-quick-reference.md
README.md
23-graph-neural-networks/
graph-neural-networks.md
graph-neural-networks-advanced-topics.md
graph-neural-networks-project-tutorial.md
graph-neural-networks-quick-reference.md
README.md
24-audio-speech-processing/
audio-speech-processing.md
audio-speech-processing-advanced-topics.md
audio-speech-processing-project-tutorial.md
audio-speech-processing-quick-reference.md
README.md
25-generative-ai-llms/
generative-ai-llms.md
generative-ai-llms-advanced-topics.md
generative-ai-llms-project-tutorial.md
generative-ai-llms-quick-reference.md
README.md
system-design/
00-computer-architecture.md
01-application-architecture.md
02-design-requirements.md
03-networking-basics.md
04-tcp-and-udp.md
05-dns.md
06-http.md
07-websockets.md
08-api-paradigms.md
09-api-design.md
10-caching.md
11-cdns.md
12-proxies-and-load-balancing.md
13-consistent-hashing.md
14-sql.md
15-nosql.md
16-replication-and-sharding.md
17-cap-theorem.md
18-object-storage.md
19-message-queues.md
20-mapreduce.md
21-designing-systems.md
22-design-rate-limiter.md
23-design-social-feed.md
24-design-group-chat.md
25-design-video-streaming.md
26-design-cloud-file-storage.md
27-design-maps-and-navigation.md
28-design-key-value-store.md
29-design-distributed-message-queue.md
30-design-realtime-feature-store.md
README.md
examples/
full_stack_track/
README.md (learner scratch space for the Full-Stack AI track)
tools/
check_links.py
README.md
resources/
agile_data_science.md
ai_agents_guide.md
automl_basics.md
blogs_websites.md
books.md
causal_inference_guide.md
career_portfolio.md
career_roadmap_guide.md
common_errors.md
courses.md
data_products_guide.md
data_science_cheatsheet.md
data_validation.md
datasets.md
docker_tutorial.md
dsa_course_python.md
dsa_for_ml_guide.md
enterprise_data_tools.md
ethics_in_ml.md
excel_data_analysis_guide.md
genai_production_deployment.md
generative_ai_comprehensive_guide.md
git_guide.md
imbalanced_data_cheatsheet.md
interview_prep.md
introduction_to_ml_cheatsheet.md
java_for_data_science.md
kaggle_competitions.md
langchain_guide.md
llamaindex_guide.md
math_formulas.md
ml_glossary.md
ai_engineering_glossary.md
ai_myths_busted.md
backend_engineering_roadmap.md
ml_model_testing.md
ml_system_design_guide.md
mlflow_comprehensive_guide.md
mlops_cheatsheet.md
model_deployment_cheatsheet.md
model_explainability_cheatsheet.md
model_interpretability.md
open_source_contribution.md
powerbi_guide.md
practice_platforms.md
prerequisites_cheatsheet.md
rag_comprehensive_guide.md
recommender_systems.md
reinforcement_learning.md
stakeholder_communication.md
tools.md
transformer_fine_tuning_guide.md
web_scraping_guide.md
youtube_channels.md
requirements.txt
LICENSE
README.md
CONTRIBUTING.md
DISCLAIMER.md
START-HERE.md
WRITING-STYLE.md
GETTING_STARTED.md
LEARNING_ROADMAP.md
QUICK_START.md
Note: Module folders are numbered 00–25 on disk. Stages in the learning path reorder topics for teaching. They are not the same label. See START-HERE.md.
- Hands-On Machine Learning by Aurélien Géron
- Pattern Recognition and Machine Learning by Christopher Bishop
- Deep Learning by Ian Goodfellow, Yoshua Bengio, and Aaron Courville
- Introduction to Statistical Learning by James et al.
- Machine Learning by Andrew Ng (Coursera)
- Deep Learning Specialization by Andrew Ng (Coursera)
- Fast.ai Practical Deep Learning
- CS229: Machine Learning (Stanford)
| Guide | Description |
|---|---|
| Data Science & ML Cheatsheet | Quick reference for daily work (NumPy, Pandas, PyTorch, TensorFlow, OpenCV, FastAPI, and more) |
| Prerequisites Cheatsheet | Quick reference for Python, Mathematics, and Statistics fundamentals needed for ML |
| Introduction to ML Cheatsheet | Quick reference for ML fundamentals, types, workflow, and key concepts |
| Math Formulas Reference | Essential mathematical formulas for ML (Statistics, Linear Algebra, Calculus, ML metrics) |
| ML Glossary | Comprehensive glossary of ML terms and concepts for beginners |
| AI Engineering Glossary | Practical LLM / agent / RAG / serving jargon: what people say vs what it means |
| AI Myths Busted | Expert myth-busting for AI/ML/DL: what people believe, what is actually going on, and one line to remember |
| Backend Engineer Roadmap | Learn backend from HTTP through production: auth, Postgres, caching, queues, security, and scaling |
| Common Errors & Debugging Guide | Troubleshooting guide for common ML errors and debugging strategies |
| DSA for ML Guide | Essential data structures and algorithms for machine learning |
| DSA Course (Python) | Full beginner-to-interview DSA course in Python (patterns, trees, graphs, practice plan) |
| Guide | Description |
|---|---|
| Complete Git & GitHub Guide | Comprehensive Git tutorial with commands, outputs, practice exercises, and solutions |
| Docker Complete Tutorial | Comprehensive Docker guide for ML: containerization, Dockerfile, docker-compose, best practices, and deployment |
| Web Scraping Guide | Complete web scraping guide from basics to advanced: Requests, Beautiful Soup, Selenium, Scrapy, CAPTCHA handling |
| Backend Engineer Roadmap | HTTP, REST, Postgres, caching, queues, security, scaling, and ops with free links and repo cross-refs |
| Full-Stack Track (in-repo lessons) | TypeScript through AI integration: chapter lessons paired with the Full-Stack AI Blueprint |
| Guide | Description |
|---|---|
| MLFlow Comprehensive Guide | Complete MLFlow guide: experiment tracking, model registry, hyperparameter tuning, MLFlow UI, Docker deployment |
| MLOps Cheatsheet | Quick reference for MLOps tools, practices, and workflows (DVC, MLflow, CI/CD, monitoring) |
| Model Deployment Cheatsheet | Quick reference for deploying ML models (APIs, Docker, cloud platforms, A/B testing) |
| Imbalanced Data Cheatsheet | Quick reference for handling imbalanced datasets (resampling, class weights, metrics) |
| Model Explainability Cheatsheet | Quick reference for explaining ML models (SHAP, LIME, feature importance, PDP) |
| Transformer Fine-Tuning Guide | Comprehensive guide to fine-tuning transformers (T5, BERT, GPT) with Hugging Face |
| Model Interpretability Guide | Understanding and explaining ML model predictions (SHAP, LIME, feature importance) |
| Reinforcement Learning Basics | Introduction to RL, key concepts, algorithms, and applications |
| Recommender Systems Guide | Building recommendation systems (collaborative filtering, content-based, hybrid approaches) |
| AutoML Basics Guide | Introduction to Automated Machine Learning: when to use, popular tools, and integration strategies |
| Data Validation Guide | Comprehensive data validation: schema validation, quality checks, drift detection, and automated pipelines |
| Causal Inference Guide | Comprehensive guide to causal inference: potential outcomes, confounding, RCTs, observational methods (propensity scores, DiD, IV, RDD), causal ML, and tools (DoWhy, EconML) |
| Guide | Description |
|---|---|
| RAG Comprehensive Guide | Complete guide to Retrieval Augmented Generation: architecture, components, vector databases, advanced techniques, evaluation, and production deployment |
| Langchain Guide | Complete Langchain guide (Python-first examples; JavaScript / TypeScript docs linked at the top for Node and Next.js stacks): chains, agents, memory, RAG, document loaders, and vector stores |
| LlamaIndex Guide | Comprehensive LlamaIndex guide: data indexing, querying, retrieval, chat engines, and advanced generative AI projects |
| AI Agents Guide | Complete guide to AI agents: CrewAI, AutoGen, Langgraph, AutoGPT, MCP (Model Context Protocol), and A2A (Agent-to-Agent) communication |
| GenAI Production Deployment Guide | Comprehensive guide to deploying GenAI at scale: RAG architectures, multi-agent systems, hyperscaler deployment (AWS, GCP, Azure), scaling strategies, monitoring, and cost optimization |
| Generative AI Comprehensive Guide | Complete overview of Generative AI: LLMs, LangChain, RAG, AI Agents, Vector Databases, Multi-Agent Systems, and building production-ready GenAI applications |
| AI Engineering Glossary | Fast definitions for production GenAI terms (agents, RAG, KV cache, LoRA, evals, guardrails) |
| AI Myths Busted | Busts fluency, scale, RAG, agents, eval, and career myths with sticky one-liners |
| Guide | Description |
|---|---|
| System Design for Beginners | 22 foundation lessons (computer architecture, networking, HTTP, WebSockets, API design, caching, CDNs, proxies, load balancing, consistent hashing, SQL/NoSQL, replication, sharding, CAP, object storage, message queues, MapReduce) + a capstone + 9 interview-style designs (rate limiter, social feed, group chat, video streaming, cloud file storage, maps, distributed KV store, distributed message queue, real-time ML feature store). Read this before the ML guide below |
| ML System Design Guide | Applies system design fundamentals to ML services: requests/responses, latency, throughput, load balancing, caching, vertical/horizontal scaling, databases, replication, sharding, message queues, stateless/stateful architecture, high availability, monitoring, drift, MLOps |
| Guide | Description |
|---|---|
| Stakeholder Communication Guide | Effective communication of ML concepts, results, and business value to non-technical stakeholders |
| ML Model Testing Guide | Comprehensive guide to testing ML models, pipelines, and APIs (unit tests, integration tests, best practices) |
| Ethics in ML Guide | Comprehensive guide to bias, fairness, responsible AI, and ethical ML practices |
| Agile Data Science Guide | Applying Agile methodologies (Scrum, Kanban) to data science projects: sprint planning, standups, retrospectives |
| Data Products Guide | Building production-ready data products: APIs, dashboards, recommendation systems, real-time analytics |
| Enterprise Data Tools Guide | Enterprise data platforms: Snowflake, Informatica, Talend, Cloudera, Stibo, Qlik, Tableau integration |
| Java for Data Science Guide | Using Java in data science: ML libraries (Weka, Deeplearning4j), Spark integration, enterprise systems |
| Guide | Description |
|---|---|
| Power BI Guide | Complete Power BI guide: visualizations, DAX, Power Query, data modeling, dashboards, and AI integration |
| Excel Data Analysis Guide | Comprehensive Excel guide: functions, pivot tables, charts, dashboards, Power Query, and advanced techniques |
Follow my curated playlist for guided video walkthroughs:
- Career Roadmap Guide: Role-specific learning paths: Data Analyst, Data Scientist, ML Engineer, LLM Engineer, GenAI Solution Architect, Computer Vision Engineer, AI Engineer, Data Engineer, MLOps Engineer, Research Scientist, BI Analyst
- Full-Stack AI Engineer (overview) · In-repo lessons · Blueprint + free resources: TypeScript through deployment and AI product integration
- Career & Portfolio Guide: Build your portfolio, prepare for interviews, and launch your ML career
- Interview Preparation Guide: ML interview questions, coding challenges, system design, and preparation strategies
- Open Source Contribution Guide: How to contribute to open source projects in data science and ML
- Kaggle Competitions Guide: Complete guide to participating in Kaggle competitions, strategy, and collaboration
- Docker Complete Tutorial: Comprehensive Docker guide for ML: containerization, Dockerfile, docker-compose, best practices, and deployment
- Books: Recommended books for ML and data science
- Courses: Online courses and learning platforms
- Datasets: Curated list of datasets for practice
- Tools: Essential tools and libraries for ML
- YouTube Channels: Comprehensive list of free ML YouTube channels
- Blogs & Websites: Recommended blogs, websites, and online resources
- Practice Platforms: Platforms to practice ML, coding, and data science
- Follow the core path first: Modules build on each other. Finish Stage 0–2 before jumping to deep learning unless you already know the basics.
- Practice Regularly: Code along with examples and complete all exercises
- Build Projects: Apply what you learn by building projects
- Join Communities: Engage with others learning ML
- Be Patient: ML is complex - take your time to understand concepts
- Experiment: Don't just copy code - experiment and break things!
Note: These are common questions that learners typically have when starting their ML journey. They're based on typical learning patterns and common concerns, not necessarily from actual user submissions.
Q: I'm a complete beginner. Where do I start?
A: Read START-HERE.md, then open 00-prerequisites/01-python-basics.md. No prior experience needed. Follow modules in order.
Q: How long will it take to complete this roadmap?
A: Realistic time estimates vary significantly based on your background and time commitment:
| Commitment Level | Hours/Week | Minimum | Standard | Comprehensive |
|---|---|---|---|---|
| Full-Time | 30-40 hrs | 12-15 months | 15-18 months | 18-22 months |
| Part-Time | 10-15 hrs | 24-30 months | 30-36 months | 36-39 months |
Definitions:
- Minimum: Core concepts, skips some advanced topics
- Standard: Complete coverage with all modules and projects
- Comprehensive: Deep understanding, all projects, portfolio building
| Stage | Focus | Full-Time | Part-Time |
|---|---|---|---|
| Prerequisites | Python, Math | 2-3 months | 4-6 months |
| Data Fundamentals | NumPy, Pandas, Visualization | 2-3 months | 4-6 months |
| ML Basics | Regression, Classification, Evaluation | 2-3 months | 4-6 months |
| Advanced ML | Ensembles, Feature Engineering | 1-2 months | 2-4 months |
| Deep Learning Fundamentals | Neural Networks, Frameworks | 2 months | 4 months |
| Specialized Deep Learning | Computer Vision, NLP, Time Series | 3-4 months | 6-8 months |
| Essential Skills | SQL, Imbalanced Data, Explainability | 1-2 months | 2-4 months |
| Production & MLOps | Deployment, MLOps | 2-3 months | 4-6 months |
| Projects (23 total) | Hands-on Practice | 4-6 months | 8-12 months |
| Total | Complete Path | 15–22 months | 30–39 months |
| Factor | Impact on Timeline |
|---|---|
| Prior programming experience | -2 to -4 months |
| Prior math background | -1 to -2 months |
| Number of projects completed | +2 to +6 months |
| Practice vs. reading ratio | Practice takes longer but is essential |
Q: Do I need a powerful computer?
A: No! Most modules work fine on a regular laptop. Deep learning modules (11-12) benefit from GPUs but can be done on cloud platforms (Google Colab, Kaggle) for free.
Q: Should I learn X before Y?
A: Generally, follow the module order. However:
- SQL (module 19) can be learned after Stage 1 (Data Fundamentals)
- Imbalanced Data (module 20) is best after Classification (module 4)
- Model Explainability (module 21) is best after Advanced ML (Stage 3)
Q: Can I skip modules?
A: We recommend following the order, but you can:
- Skip advanced topics files if you're pressed for time (come back later)
- Learn SQL earlier if you need it for data access
- Jump to projects relevant to your goals
Q: I'm a software engineer. Can I skip Python basics?
A: Review 00-prerequisites/01-python-basics.md quickly - it includes ML-specific Python concepts (time complexity, iterators, generators) that may be new.
Q: I'm a statistician. Can I skip the math modules?
A: Review them quickly. They focus on ML applications of math concepts you may know from a different angle.
Q: What if I get stuck on a concept?
A:
- Re-read the explanation
- Check the "Additional Resources" section in module READMEs
- Look for related topics in other modules
- Practice with code examples
- Join communities (see Resources section)
Q: Which IDE should I use?
A: Any works! Popular choices:
- Jupyter Notebooks: Great for learning and experimentation
- VS Code: Excellent for larger projects, good ML extensions
- PyCharm: Full-featured Python IDE
- Google Colab: Free cloud-based notebooks
Q: Do I need to install everything at once?
A: No! Install packages as you need them. Each module lists required packages. Start with requirements.txt for core packages.
Q: Python 3.8, 3.9, 3.10, or 3.11?
A: Python 3.10–3.12 recommended (3.10 or 3.11 for widest library compatibility). Use a virtual environment and pip install -r requirements.txt from the repository root.
Q: Should I do all projects?
A: Do at least:
- 2-3 beginner projects (after Stage 2)
- 3-4 intermediate projects (after Stage 4)
- 2-3 advanced projects (after Stage 8) Focus on projects relevant to your career goals.
Q: Can I use my own datasets?
A: Absolutely! Using your own data makes projects more meaningful. Just ensure the dataset is appropriate for the project type.
Q: How long should each project take?
A: Recommended project completion by level:
| Level | Projects to Complete | Time per Project | Total Time |
|---|---|---|---|
| Beginner | 3-4 projects | 1-3 days | 1-2 weeks |
| Intermediate | 4-5 projects | 3-7 days | 3-5 weeks |
| Advanced | 2-3 projects | 1-2 weeks | 2-6 weeks |
| Total Minimum | 9-12 projects | - | 6-13 weeks |
| Recommended | 15-18 projects | - | 10-20 weeks |
Q: What jobs can I get after completing this?
A: Different roles require different module focuses:
| Role | Key Modules | Focus Areas | Est. Time |
|---|---|---|---|
| Data Scientist | 0-8, 15, 19-21 | Data analysis, modeling, SQL, explainability | 10-14 months |
| ML Engineer | 0-14 | Full stack: modeling to deployment, MLOps | 12-18 months |
| Research Scientist | 0-12, advanced topics | Deep learning, research, publications | 15-20 months |
| Business Analyst | 0-7, 19 | Data analysis, SQL, business context | 8-12 months |
| Data Engineer | 0-1, 13-14, 19 | Data pipelines, infrastructure, SQL | 10-14 months |
Q: Do I need a degree?
A, not necessarily! Many successful ML practitioners are self-taught. However, a degree can help with:
- Getting past HR filters
- Research positions
- Certain companies' requirements
Q: Should I learn TensorFlow or PyTorch?
A: Both! Start with TensorFlow/Keras (easier for beginners), then learn PyTorch. Many jobs use both. Module 10 covers both.
Q: Do I need to know deep learning for most ML jobs?
A, not always! Many roles focus on traditional ML (modules 3-8). However, deep learning (modules 9-12) is increasingly important.
Q: How important is MLOps?
A: Very important for production ML! Module 14 covers MLOps basics. Essential for ML Engineer roles, valuable for Data Scientists too.
Q: Should I learn SQL?
A: Yes! Most data science roles require SQL. Module 19 covers it comprehensively. Learn it early if you need data access skills.
Q: I'm overwhelmed. What should I do?
A:
- Take a break
- Focus on one module at a time
- Don't try to master everything immediately
- Practice regularly (even 30 min/day helps)
- Join study groups or communities
Q: I don't understand the math. Should I continue?
A:
- Review the math modules (00-prerequisites) with visual resources (3Blue1Brown videos)
- Focus on intuition over proofs initially
- Use code to understand concepts
- Math becomes clearer as you apply it
Q: My code doesn't work. What should I do?
A:
- Read error messages carefully
- Check
resources/common_errors.md - Verify you're using correct library versions
- Search Stack Overflow with the error message
- Check module "Common Issues" sections
Q: How do I know if I'm ready for the next module?
A: You're ready when you can:
- Explain the main concepts
- Implement basic examples without copying
- Complete exercises (even if with some help)
- Understand most of the code examples
Q: How can I contribute?
A, see CONTRIBUTING.md. We welcome:
- Fixing typos and errors
- Adding examples
- Improving explanations
- Adding projects
- Translating content
Q: Can I use this content for my course/tutorial?
A: Yes! This is open source (MIT License). Please credit the repository and contributors.
Have more questions? Open an issue on GitHub or check the module-specific README files for detailed information!
We welcome contributions! This repository is for the community, by the community. Here's how you can help:
- Add Projects: Share your ML projects with the community
- Improve Documentation: Fix typos, clarify explanations, add examples
- Create Exercises: Add practice problems and solutions
- Report Issues: Found a bug? Let us know!
- Suggest Features: Have an idea? Open an issue!
- Fork the repository
- Clone your fork:
git clone https://github.com/YOUR_USERNAME/road-to-machine-learning.git - Create a branch:
git checkout -b feature/amazing-feature - Make your changes
- Commit:
git commit -m 'Add amazing feature' - Push:
git push origin feature/amazing-feature - Open a Pull Request
See CONTRIBUTING.md for detailed guidelines.
Your contributions make this resource better for everyone!
This project is licensed under the MIT License. See the LICENSE file for details.
If you find this repository helpful, please consider:
- Star this repo - It helps others discover this resource
- Fork it - Create your own learning path
- Share it - Help others on their ML journey
- Contribute - Add projects, fix issues, improve content
Every star and fork helps the community grow!
By completing this roadmap, you'll be able to:
- Build and train ML models from scratch
- Deploy models to production
- Work with real-world datasets
- Understand deep learning concepts
- Create computer vision and NLP applications
- Implement MLOps best practices
- Build a portfolio of ML projects
| Metric | Count | Details |
|---|---|---|
| Learning Modules | 26 | Modules 00-24, 25 covering all ML topics from basics to advanced including Generative AI |
| Projects | 23 | 6 beginner + 8 intermediate + 9 advanced with complete code and READMEs |
| Resource Guides | 50 | Cheatsheets, tutorials, and career guides |
| Markdown Files | 270+ | Lessons, code examples, and exercises |
| Learning Time (Full-Time) | 15-22 months | 30-40 hours/week for comprehensive coverage |
| Learning Time (Part-Time) | 30-39 months | 10-15 hours/week for comprehensive coverage |
| Prerequisites | Zero | Start from scratch. No prior experience needed! |
External Links Disclaimer:
This repository contains links to external websites, courses, documentation, and resources provided for educational purposes only. The maintainers:
- Do not endorse any specific external service or content provider
- Are not responsible for the availability, accuracy, or content of external links
- Cannot guarantee that external links will remain accessible or unchanged
- Do not assume liability for any issues arising from the use of external resources
GDPR Compliance:
This repository does not collect, store, or process any personal data. It is a static educational resource. Any data processing occurs through GitHub (the platform) or external websites, which have their own privacy policies. See DISCLAIMER.md for complete disclaimer and GDPR information.
Made with ❤️ for the ML community
Happy Learning!
Remember: The journey of a thousand miles begins with a single step. Start with module 00 and keep going!
