AI Engineering

Weeks 9–14 · 8 Modules. Go beyond prompts. This is where VWC graduates separate from every other bootcamp — you won't just use AI tools, you'll build the systems behind them. Production APIs, retrieval-augmented generation, autonomous agents, and real-time streaming interfaces.

Module 14: AI Foundations — Before you write AI code, understand what AI actually is. Covers the AI/ML/Deep Learning hierarchy, learning paradigms, the Transformer architecture, tokens and embeddings, context windows, and the LLM landscape (GPT-4, Claude, Gemini, Llama).

Module 15: Advanced Python for AI — Level up to production patterns. Covers Python 3.11+ features, uv, type hints, Pydantic v2 (BaseModel, validators, serialization, BaseSettings, computed fields, JSON schema generation), and async Python (asyncio, httpx).

Module 16: FastAPI: Production AI APIs — Build the backend that powers AI applications at scale. Covers routing, Pydantic request/response, dependency injection, OpenAPI 3.1, streaming responses (SSE, AsyncIterator), middleware, CORS, API key auth, rate limiting, and background tasks.

Module 17: Google Gemini Integration — Connect to frontier AI models for text and multimodal applications. Covers Gemini Pro/Flash, generation config, text generation, streaming, chat sessions, structured output (JSON mode + Pydantic schema), multimodal inputs, and function calling.

Module 18: Professional Prompt Engineering — Systematic techniques that consistently produce better results. Covers zero/one/few-shot, Chain of Thought, Self-Consistency, Tree of Thoughts, ReAct, XML tags, schema-guided generation, Jinja2 templates, and A/B testing.

Module 19: RAG Systems — Give AI applications real knowledge. Covers document ingestion, chunking strategies, embeddings, pgvector and ChromaDB, hybrid search, reranking, query expansion (HyDE), and evaluation metrics (Precision@k, Recall@k, MRR, faithfulness).

Module 20: AI Agents & Workflows — Build autonomous systems that reason, plan, and act. Covers agent architecture (perception, reasoning, action, reflection), tool use, function calling, memory systems, LangChain, LangGraph, multi-agent systems, and human-in-the-loop patterns.

Module 21: Full-Stack AI Integration — Connect everything. Covers Next.js AI patterns, Server Components for AI, streaming chat UI (SSE, token-by-token rendering, markdown), type-safe integration (OpenAPI client generation, Zod ↔ Pydantic), and file upload with multimodal support.

Phase Outcome: Design, build, and deploy AI-powered applications from scratch.