Overview
Running a language model in a demo is easy; running it inside a product, with real users, measurably and safely is a different job. In this training we connect an LLM API to an application, let the model call your own tools, ground its answers in your data and measure quality with automated evaluation. How personal data and trade secrets reach the model is part of every session.
Course content
1. Language models
- How a model answers: tokens and the context window
- Choosing a model: capability, latency, cost
- What the model does not know: recency and your data
2. Using the API
- Messages and system instructions
- Streaming
- Prompt caching
- Handling errors and rate limits
3. Writing instructions
- Clear tasks and examples
- Structured output with JSON schemas
- Versioning prompts
- Tool definitions and schemas
- Multi-step tool loops
- Limiting tool permissions
5. Answering from your data (RAG)
- Chunking documents
- Embeddings and vector search
- Hybrid search and re-ranking
- Citations
- Lab: answers from company documents with sources
6. Evaluation
- Building a test set
- Automated grading and model-graded evals
- Regression tracking
7. Security
- Prompt injection
- Treating external content as data
- Validating output
8. Data protection
- What reaches the model and where it is processed
- Personal data and cross-border transfer under KVKK
- Masking and data retention settings
- Measuring token cost
- Smaller models and caching
- Monitoring latency
10. Agent workflows
- When an agent helps and when it is too much
- Human approval and reversibility
- Exposing tools with MCP
- Lab: an evaluation set and a guard layer
Schedule
Each module takes about 4 hours, half a day; 40 hours in total.
- Day 1: modules 1 and 2
- Day 2: modules 3 and 4
- Day 3: modules 5 and 6
- Day 4: modules 7 and 8
- Day 5: modules 9 and 10
Prerequisites
- Experience in a programming language (C#, TypeScript or Python)
- Having worked with HTTP APIs
- No machine learning background needed
- On site or remote: at your office or in live virtual sessions.
- Hands-on: every day ends with lab sessions where you apply that day’s topic to a working example.
- Lab environment: participants get a prepared working environment and a sample code repository.
- Tailored to your team: content and examples are adapted to your stack after a short call before the training.
- Materials: slides, lab guides and reference code stay with you after the training.