All courses
33 published courses. Narrow by subject, role, level or format, or search by the outcome you are after.
Filter1
Filter courses
DoneFor your role
Competency
Format
Matching courses
advancedTechnical AI
LLM-as-Judge Lab: Calibration and DriftA hands-on workshop for data analysts who already call LLM APIs and know eval-framework basics.
advancedTechnical AI
Building and Securing a Custom MCP ServerThis advanced workshop teaches AI engineers to build a production-grade custom MCP server with the official SDKs and secure it end to end: a scoped OAuth 2.1 authorization handshake, input validation and sandboxing, TLS deployment with rate limiting and connection quotas, audit logging, and an ad…
advancedTechnical AI
Retrieval Quality Lab: Chunking, Reranking, EvaluationA hands-on workshop for data analysts who already know embeddings and vector search, teaching them to build, measure and improve a retrieval pipeline.
advancedTechnical AI
Multi-Agent Orchestration LabA hands-on workshop for engineers who have already shipped a single-agent LLM application and want to design, build, trace and evaluate systems where three or more specialized agents coordinate through a shared protocol.
intermediateTechnical AI
RAG Over Your Own DocumentsA hands-on course for data analysts who already call LLM APIs and understand embeddings conceptually.
intermediateTechnical AI
Building an Eval HarnessFor AI-savvy data analysts moving from ad hoc prompt testing to a real evaluation practice, this course builds a working eval harness end to end: a labeled test dataset, deterministic and LLM-as-judge graders, pass/fail thresholds tied to real stakes, regression detection, and a CI-wired scorecard.
advancedTechnical AI
Prompt Engineering Lab: Systematic Iteration and RegressionThis workshop-format course teaches AI engineers who already iterate on prompts how to treat prompts as versioned, tested software artifacts rather than one-off tweaks.
intermediateTechnical AI
Building Your First Production AgentA hands-on course for engineers who already call LLM APIs and want to ship a tool-using agent that survives production: recovering from failed tool calls, capping runaway spend, extending itself with an MCP server, passing task-completion evals, and deploying with logging enabled.