Courses › Course07

Course07 · AI Engineering: Agents and Applications

Agents, MCP, AI coding tools, evaluation, and full-stack AI applications — engineering AI systems that hold up in production.

42 days 42 written 42 complete

Days 288-294 · AI Agents 7/7 complete
  1. Day 288 What an AI Agent Is complete
  2. Day 289 The Agent Loop: Reason, Act, Observe complete
  3. Day 290 Designing Tools for Agents complete
  4. Day 291 An Agent from Scratch complete
  5. Day 292 Agent Frameworks and When to Use Them complete
  6. Day 293 Multi-Agent Systems complete
  7. Day 294 Building a Research Agent complete

Project: Research Agent — Build an agent that researches a question with search and note-taking tools and produces a cited brief.

Days 295-301 · The Model Context Protocol 7/7 complete
  1. Day 295 What MCP Is and Why It Exists complete
  2. Day 296 Using MCP Servers complete
  3. Day 297 Building an MCP Server complete
  4. Day 298 MCP Resources and Prompts complete
  5. Day 299 Building an MCP Client complete
  6. Day 300 MCP Security complete
  7. Day 301 Your Personal MCP Server complete

Project: Personal MCP Server — Build and test an MCP server exposing your own data or tools, and use it from a real client.

Days 302-308 · AI Coding Agents 7/7 complete
  1. Day 302 The AI Coding Landscape complete
  2. Day 303 Working with a Coding Agent complete
  3. Day 304 Effective Agentic Coding Workflows complete
  4. Day 305 Configuring Agents: Memory, Skills, and Rules complete
  5. Day 306 Reviewing and Trusting AI-Written Code complete
  6. Day 307 Coding Agents in CI and Automation complete
  7. Day 308 Shipping a Feature with an Agent complete

Project: Agent-Built Feature — Ship a reviewed, tested feature to one of your own projects using an AI coding agent, documenting the workflow.

Days 309-315 · Evaluation and Reliability 7/7 complete
  1. Day 309 Why Evals Are the Real Moat complete
  2. Day 310 Building Evaluation Datasets complete
  3. Day 311 LLM-as-Judge complete
  4. Day 312 Regression Testing for Prompts and Models complete
  5. Day 313 Guardrails and Content Moderation complete
  6. Day 314 Observability and Tracing for AI complete
  7. Day 315 An Evaluation Harness complete

Project: Evaluation Harness — Build an eval harness with a labeled dataset, an LLM judge, and regression gates for one of your AI features.

Days 316-322 · Full-Stack AI Applications 7/7 complete
  1. Day 316 Architecture of an AI Product complete
  2. Day 317 Backend Patterns for LLM Apps complete
  3. Day 318 Chat UX and Streaming Frontends complete
  4. Day 319 Auth, Quotas, and Billing complete
  5. Day 320 Latency and Caching complete
  6. Day 321 Vendor Abstraction and Fallbacks complete
  7. Day 322 A Full-Stack AI Application complete

Project: Full-Stack AI App — Build a deployed chat application with streaming, auth, usage quotas, and a vendor-abstraction layer.

Days 323-329 · Production Retrieval and Pipelines 7/7 complete
  1. Day 323 Data Ingestion Pipelines complete
  2. Day 324 Document Processing at Scale complete
  3. Day 325 Keeping Indexes Fresh complete
  4. Day 326 Scaling Retrieval complete
  5. Day 327 Cost Engineering for AI Systems complete
  6. Day 328 Privacy in AI Systems complete
  7. Day 329 Section Project: A Production Assistant complete

Project: Section Project: Production Assistant — Assemble a production-grade assistant: ingestion pipeline, fresh index, evals, cost budget, and privacy review.