From Zero to ML Engineer: A 6-Month Roadmap

“Learn AI in 30 days” rarely produces someone who can own a training pipeline or debug a serving outage. This roadmap spreads the work across six months with verifiable milestones.

Months 1–2: Foundations

• Python, NumPy, pandas, basic linear algebra (vectors, matrices, dot products).
• Build: train logistic regression on a tabular dataset; explain train/val split and overfitting in your own words.

Month 3: Deep learning core

• PyTorch tensors, autograd, a simple MLP on MNIST or CIFAR-10.
• Read alongside: how transformers work once you are comfortable with backprop.

Month 4: NLP / LLM basics

• Tokenization, embeddings, calling APIs, building a small RAG app.
• Study fine-tuning vs RAG before spending money on GPUs.

Month 5: Agents and deployment

• Follow building your first AI agent with logging and tests.
• Containerize a FastAPI inference or RAG service; deploy to a cloud free tier.

Month 6: Portfolio and interviews

• Ship two public projects—ideas in 5 AI portfolio projects that get you hired.
• Write READMEs with architecture diagrams and failure modes you handled.

How to know you are on track

Each month should end with something runnable on GitHub and a short blog-style write-up. Employers hire evidence, not course completion badges alone.