Skip to main content
Univelcity
EngineeringIntermediate

AI Engineering

Knowing how to call a model is not the same as shipping something that works in production. AI engineering is the craft of turning language models into reliable software — retrieval, tools, memory, evaluation, and the plumbing that holds it together under real load.

Who it's for

  • Developers who want to build production AI applications, not demos.
  • Engineers moving from traditional software into LLM-powered systems.
  • Vibe coders ready to go deeper and own the whole build.
  • Technical founders building an AI product themselves.

Prerequisite: you can program in at least one language and are comfortable with APIs.

What you'll build

A deployed, production-grade LLM application — with RAG, tools, and an evaluation harness — running live, with the repo to back it.

What you'll learn

  • Build production applications on top of large language models, end to end.
  • Implement retrieval-augmented generation (RAG) with embeddings and a vector store.
  • Give models tools and let them take actions safely.
  • Add evaluation, observability, and guardrails so your system stays reliable.
  • Ship, deploy, and iterate on a real LLM application.
  1. 01
    Foundations of building with LLMs
    How models behave, the API, structured output, and the anatomy of a production AI app.
  2. 02
    Prompt engineering that survives production
    Reliable prompting, structured output, and controlling model behaviour.
  3. 03
    Retrieval-augmented generation
    Embeddings, chunking, vector stores, and building a RAG pipeline over your own data.
  4. 04
    Tools and actions
    Function calling, giving models tools, and letting systems do real work safely.
  5. 05
    Memory, state, and orchestration
    Managing context, conversation, and multi-step workflows.
  6. 06
    Evaluation and reliability
    Measuring quality, catching regressions, observability, and guardrails.
  7. 07
    Cost, latency, and scale
    Caching, model choice, and the economics of inference.
  8. 08
    Ship it
    Deploy, monitor, and iterate on a live application. Capstone.
FORMAT

Live & hands-on

Live, expert-led, small cohort · 8 weeks · online · hands-on.
WALK AWAY

Certificate + proof of work

Univelcity AI Engineering certificate + a live production AI application and repo.
TAUGHT BY

Practitioners

TODO — instructor details.

Questions

Good to know.

How much coding experience do I need?
You should already program and be comfortable with APIs. If you don't code yet, start with Vibe Coding.
Which model and stack?
You'll build with a current LLM API (including Claude) and the tools used on production AI teams. Principles transfer across providers.
Does this prepare me for a Claude Developer certification?
It maps closely — see the Certification track for focused exam prep.

Build AI that survives production.

Apply for the next cohort.

Apply