Generative AI Engineering
A 12-week live cohort on building with large language models: retrieval, tool use, evaluation, cost control and the safety work that separates a demo from a product.
Who this track is for
Developers who can already build an application and want to add language-model features that survive contact with real users.
Prerequisites: Working knowledge of Python or JavaScript and comfort calling an HTTP API. No machine-learning background needed.
What you will learn
How language models behave
Tokens, context windows, sampling and a practical mental model of what these systems can and cannot do.
Prompting as engineering
Structured prompts, output schemas, few-shot design and version-controlling prompts like code.
Retrieval-augmented generation
Chunking, embeddings, vector search, re-ranking and grounding answers in sources you control.
Tools and agents
Function calling, multi-step workflows, state, retries and knowing when an agent is the wrong shape.
Evaluation
Building a test set, automated and human evaluation, regression testing a non-deterministic system.
Cost, latency and safety
Caching, streaming, model selection, rate limits, prompt injection, PII handling and abuse controls.
What you finish with
The point of the track is the evidence, not the certificate. By the end you will have built and deployed:
- A retrieval-grounded assistant over a document set
- A tool-using workflow with structured, validated output
- An evaluation harness with a scored regression suite
- A cost- and latency-instrumented deployment
Plus a Medhavio certificate of completion. That certificate is issued by us; it is not a university or government accreditation, and the portfolio above is the part an interviewer will actually look at.
Tools you will work in
- LLMs
- RAG
- Agents
Product and technology names are trademarks of their respective owners and are listed only to describe what the track teaches. Naming them implies no affiliation or endorsement.
How the cohort runs
- Weekly live sessions in a small group, recorded so you can catch up.
- A project every module, reviewed line by line by a practising mentor.
- 1:1 mentor sessions you book when you are stuck or planning your next step.
- An AI study copilot for hints, code review and a revision plan built around your week.
- Interview preparation: mock rounds, portfolio review and negotiation practice.
- Community access that continues after the cohort ends.
Fees, inclusions and the free tier are on the pricing section. Cancellation windows are in the Refund & Cancellation Policy. We prepare you for interviews; we do not guarantee a job, a placement or a salary.