A self-paced course on building applications on top of language models. It covers six stages: what these systems can and cannot do, prompting and structured output, retrieval augmented generation, agents and tool use, evaluation and testing, and the production concerns of cost, latency, safety and monitoring.

The emphasis is on the engineering discipline around a non-deterministic component, which is what separates a demo that impresses from a system that holds up. Tech Skills Library is independent and issues no certification or credential.

Modules

Stage One — What These Systems Can and Cannot Do

Lessons

• The Component You Are Building Around Preview
• Choosing a Model and Where to Run It
• Deciding Whether the Model Belongs Here At All

Stage Two — Prompting and Structured Output

Lessons

• Prompt Construction That Holds Up
• Structured Output and Defensive Parsing
• Decomposition and Chaining

Stage Three — Retrieval Augmented Generation

Lessons

• Why Retrieval and What It Fixes
• Chunking Embedding and Indexing
• Improving Retrieval Quality

Stage Four — Agents and Tool Use

Lessons

• What an Agent Is and When It Helps
• Designing Tools the Model Can Use
• Safety Boundaries for Actions

Stage Five — Evaluation and Testing

Lessons

• Building an Evaluation Set
• Grading Non-Deterministic Output
• Regression Testing and Shipping Changes

Stage Six — Production Cost Latency and Safety

Lessons

• Cost and Latency Control
• Safety Guardrails and Prompt Injection
• Monitoring Feedback and the Career Path