All courses
AI EngineeringPractitioner5 daysIn-person · KL
Build with LLMs
A practitioner bootcamp for building production-grade LLM features.
Five days of intensive, hands-on engineering for developers who want to ship LLM-powered features that survive contact with real users. Covers evaluation, observability, retrieval, agent patterns, and the unglamorous infrastructure that determines whether your feature works.
Outcomes
What you'll leave with.
- A capstone LLM feature deployed end-to-end
- A working evaluation pipeline you can reuse
- Confidence operating LLM features in production
Prerequisites
Comfort writing Python or TypeScript. Some experience deploying web services.
Curriculum
Day by day.
- 01
Day 1 · Setting up for serious work
- Provider abstractions and when they help
- The Teragrid Ai Platform SDK
- Your first evaluation harness
- 02
Day 2 · Retrieval and tools
- Embeddings, vector stores, and chunking
- Tool use and the structured-output discipline
- When retrieval is the wrong answer
- 03
Day 3 · Agents in production
- Agent loops, planning, and termination
- Building with Teragrid Agent
- Observability and tracing
- 04
Day 4 · Reliability and cost
- Failure modes and recovery
- Caching strategies that actually work
- Cost modelling for production traffic
- 05
Day 5 · Capstone and ship
- Build and deploy a working feature
- Pair reviews with the cohort
- A graduation showcase to the AITG team
Your instructor
Maya R.
Lead Instructor · AI Engineering Track
Maya leads the engineering-track bootcamps at trainai. Her teaching focuses on the parts of AI development that are skipped in most courses — evaluation, observability, and the long tail of production failures. (Placeholder bio.)