@heynavtoor: AI engineer job postings grew ...
AI engineer job postings grew 143% year-over-year. Average salary in the US: $175,000. For every 10 open roles in India, there is one qualified candidate.
You do not need a degree. You do not need a bootcamp. Everything you need to learn this skill is free, public, and available right now.
This is the 6-month roadmap. One month per phase. Every resource is free. No filler. Just the path.
Month 1: Python and Programming Fundamentals
Nothing works without this. Every AI framework, every library, every tool is built on Python. Do not skip this. Do not rush this.
What to learn: Variables, functions, loops, conditionals, data structures (lists, dictionaries, sets), object-oriented programming, file handling, error management, Git and GitHub basics.
Resources:
Python for Everybody (Dr. Chuck, University of Michigan) - Full course, free on YouTube and Coursera. The most popular Python course ever created.
CS50P: Introduction to Programming with Python (Harvard, David Malan) - Free on YouTube. Harvard-quality, zero prerequisites.
Automate the Boring Stuff with Python (Al Sweigart) - Free to read online. Practical Python from day one.
Git and GitHub for Beginners (freeCodeCamp) - Free on YouTube. 1 hour. Covers everything you need.
Milestone: You can write a Python script that reads a CSV, processes data, and outputs results. You have a GitHub account with 3+ projects pushed.
Month 2: Mathematics and Statistics
You do not need a math degree. You need enough math to understand why models work and what to do when they do not.
What to learn: Linear algebra (vectors, matrices, dot products, eigenvalues), calculus (derivatives, gradients, chain rule), probability (Bayes theorem, distributions), statistics (mean, variance, hypothesis testing, regression).
Resources:
3Blue1Brown: Essence of Linear Algebra - Free on YouTube. 16 videos. The best visual math content ever made.
3Blue1Brown: Essence of Calculus - Free on YouTube. Same quality. Same clarity.
Khan Academy: Statistics and Probability - Free. Comprehensive. Self-paced.
MIT 18.06: Linear Algebra (Gilbert Strang) - Free on MIT OCW. The gold standard university course.
StatQuest with Josh Starmer - Free on YouTube. Statistics explained with zero jargon.
Milestone: You understand gradient descent intuitively. You can explain what a loss function does and why matrix multiplication matters for neural networks.
Month 3: Machine Learning Fundamentals
This is where it starts to feel real. Models. Training. Prediction. Evaluation.
What to learn: Supervised learning (regression, classification), unsupervised learning (clustering, dimensionality reduction), model evaluation (accuracy, precision, recall, F1),
overfitting, cross-validation, feature engineering. Libraries: scikit-learn, pandas, NumPy, matplotlib.
Resources:
Stanford CS229: Machine Learning (Andrew Ng) - Free on YouTube. The course that started the modern ML education movement. Non-negotiable.
Google Machine Learning Crash Course - Free. Interactive. Built by Google engineers.
Kaggle Learn: Intro to ML + Intermediate ML + Feature Engineering - Free micro-courses. Hands-on from the start.
fast.ai: Practical Machine Learning for Coders - Free. Top-down approach. You build before you theorize.
Milestone: You can build, train, and evaluate a classification model on a real dataset. You have 2+ ML projects on GitHub with clean README files.
Month 4: Deep Learning and Neural Networks
The architecture behind everything: image recognition, language models, speech, generation. This is where AI gets powerful.
What to learn: Neural network fundamentals (perceptrons, activation functions, backpropagation), CNNs (image tasks), RNNs and LSTMs (sequence tasks), Transformers (the architecture behind GPT, Claude, Gemini). Frameworks: PyTorch or TensorFlow.
Resources:
Stanford CS231n: CNNs for Visual Recognition - Free on YouTube. The standard deep learning course for computer vision.
Stanford CS224n: NLP with Deep Learning - Free on YouTube. The standard NLP course. Covers Transformers in depth.
MIT 6.S191: Introduction to Deep Learning - Free on YouTube. Fast-paced, updated annually, covers the latest architectures.
fast.ai: Practical Deep Learning for Coders - Free. Build real models from lesson one. PyTorch-based.
3Blue1Brown: Neural Networks - Free on YouTube. 4 videos. The clearest visual explanation of how neural networks learn.
Milestone: You can build and train a neural network in PyTorch. You understand Transformers well enough to explain self-attention to a non-technical person.
Month 5: Generative AI, LLMs, and Agents
This is where the jobs are right now. Every company hiring AI engineers in 2026 wants this skill set.
What to learn: How LLMs work (tokenization, embeddings, attention, inference), prompt engineering, RAG (Retrieval-Augmented Generation), fine-tuning, LangChain and LlamaIndex, building AI agents, vector databases (Pinecone, Weaviate, ChromaDB), API integration (OpenAI, Anthropic, Google).
Resources:
Andrej Karpathy: Let us Build GPT From Scratch - Free on YouTube. 2 hours. The single best explanation of how GPT works, by a former OpenAI researcher.
DeepLearning.AI: LangChain for LLM Application Development (Andrew Ng) - Free short course. Hands-on.
DeepLearning.AI: Building Systems with the ChatGPT API - Free short course. Production patterns for LLM apps.
Hugging Face NLP Course - Free. Covers Transformers, fine-tuning, and deployment. The best open-source NLP resource.
LlamaIndex Documentation and Tutorials - Free. The standard framework for RAG pipelines.
Milestone: You have built a RAG application that answers questions from your own documents. You have deployed at least one LLM-powered app.
Month 6: MLOps, Deployment, and Portfolio
Building a model is 20% of the job. Getting it into production, keeping it running, and proving it works is the other 80%.
What to learn: Docker and containerization, API development (FastAPI, Flask), cloud deployment (AWS, GCP, or Azure basics), CI/CD pipelines, model monitoring, MLflow for experiment tracking, evaluation frameworks, cost optimization.
Resources:
Made With ML (Goku Mohandas) - Free. The most comprehensive MLOps course available. Covers the full production pipeline.
Docker for Beginners (TechWorld with Nana) - Free on YouTube. Practical, clear, no fluff.
FastAPI Documentation and Tutorial - Free. Build production-ready APIs for your models.
MLflow Documentation and Quickstart - Free. Industry-standard experiment tracking.
Full Stack Deep Learning (UC Berkeley) - Free on YouTube. The bridge between ML research and production engineering.
Milestone: You have 3-5 end-to-end projects on GitHub. At least one is deployed and live. Your LinkedIn and portfolio clearly demonstrate AI engineering skills.
What to Build Along the Way
Month 1-2: Data analysis scripts, web scrapers, automation tools.
Month 3: Prediction model on a real dataset (housing prices, customer churn, fraud detection).
Month 4: Image classifier or sentiment analysis model trained from scratch.
Month 5: RAG chatbot that answers questions from uploaded documents. AI agent that completes multi-step tasks.
Month 6: Full-stack AI app deployed to the cloud. End-to-end pipeline with monitoring and evaluation.
Every project goes on GitHub. Every project gets a clean README. Every project gets a LinkedIn post. This is your portfolio. This is what gets you hired.
The Math
6 months. 2-3 hours per day. Zero dollars spent.
The AI job market expects 1.3 million openings in the next two years. Fewer than half will be filled. The skills gap is not closing. It is widening.
Everything on this list is free. The courses are from MIT, Stanford, Harvard, Google, and the engineers who built the models you use every day. No bootcamp will teach it better. Most will teach it worse and charge you $10,000 for the privilege.
The only thing separating you from this career is six months of focused work.
Start today. Month 1. Python. Go.