The Agentic AI Engineering Bootcamp & Certification

Understand RAG fundamentals, LangChain, embeddings, and RAG versus fine-tuning, then build a restaurant Q&A application with an end-to-end RAG pipeline and Gradio interface.

Learn the key differences between RAG and Agentic RAG and when to use each. Using Claude Code and Codex, build and compare traditional and adaptive, agent-driven RAG pipelines for a real-world restaurant use case.

Build and optimize RAG pipelines with data ingestion, embeddings, chunking, and ChromaDB. Create evaluation datasets and stress-test Agentic RAG for accuracy, response quality, reliability, and scalability.

Understand the foundations of AI evaluation, including success criteria, single-turn and multi-turn testing, grader types, and capability versus regression evaluations. Explore LLM and agent benchmarks, leaderboards, benchmark saturation, and the key tools shaping the evaluation landscape.

Build and optimize LLM and RAG evaluation pipelines using RAGAS and Claude Code. Create evaluation datasets, measure retrieval and response quality, apply LLM-as-a-judge, and access models through OpenRouter. Diagnose weaknesses and validate improvements through a hands-on project.

Build and automate AI agent evaluations using Claude Code and Arize Phoenix. Trace agent actions, assess single-turn and multi-turn performance, and measure quality, latency, and cost. Apply graders and human review to diagnose failures, detect regressions, and validate improvements through a hands-on project.

Build a real-world AI agent evaluation framework using Claude Code, Codex, and Arize Phoenix. Automate grading, trace agent behavior, measure quality and efficiency, and validate improvements through regression testing.

Build and deploy production-ready AI agents using spec-driven development, Codex, Claude Code, and the OpenAI Agents SDK. Integrate GitHub, Vercel, and Neon, automate PR reviews and testing, and manage credentials, monitoring, and error handling.

Explore open-source AI agent development, LoRA/QLoRA fine-tuning, JEV-equivalent models, and managed offerings from Anthropic and OpenAI, including deployment and operational trade-offs.

Design reliable AI agents through prompt engineering, spec-driven development, and loop and harness engineering. Optimize token usage across cost, latency, and quality, and use monitoring, tracing, evaluations, red teaming, and error analysis to improve reliability and safety.