AI Agent Engineering

Overview

A hands-on bootcamp for engineers who are done building demos. Over 6 weeks of live weekend sessions, you build two production-grade AI agents end to end — with planning, memory, tools, observability, evaluations, and real deployments. Every concept you learn gets applied immediately to one of the two projects. No isolated exercises. No throwaway code.

Stop building demos. Start building systems that get you hired.

Two projects, built from day 1

Project 1 — AI Research & Workflow Automation Agent (LangChain + LangGraph)

A stateful, multi-step agent that takes a research goal, breaks it down, searches the web, reads and summarises sources, drafts a structured report, and emails it — all autonomously. Built on LangGraph for durable, controllable workflows with human-in-the-loop checkpoints, multi-tool orchestration, reflection loops, and full trace logging.

Project 2 — Customer Support AI Agent (Google ADK)

A production-style support agent built on Google ADK that handles incoming queries, retrieves answers from a company knowledge base, escalates to a human when uncertain, remembers context across conversations, and logs every interaction to a monitoring dashboard. Demonstrates Google's agent framework, tool use, memory, and multi-turn orchestration.

What you will learn

• Agentic design patterns — tool use, planning (ReAct), routing, reflection, parallelization, multi-agent collaboration, human-in-the-loop
• Clean agentic code — modular architecture, typed schemas, config-driven design, error handling
• Observability & monitoring — centralized logs, distributed traces, latency and cost metrics, live dashboards (Langfuse)
• Evaluations — golden test sets, automated eval pipelines, quality tracking over time
• Deployment — FastAPI + hosted endpoints, both projects live by Week 4

The schedule

2 sessions per weekend · 2.5 hours each · both projects run in parallel throughout.

Week 1 — Foundations & Scaffolding

Agent architecture fundamentals. Both project repos set up with base agent loop, first tools, config, and logging from day one.

Week 2 — Planning, Routing & RAG

ReAct planning and multi-tool orchestration for the automation agent. Full RAG pipeline (ingest, chunk, index, retrieve) for the assistant.

Week 3 — Memory, Reflection & Evals

Short and long-term memory. Self-critique and reflection loops. Automated evaluations and quality dashboards wired into both projects.

Week 4 — Multi-Agent, HITL & Production

Multi-agent collaboration (planner, executor, reviewer). Human-in-the-loop approval gates. Full observability stack live. Both projects deployed, documented, and interview-ready.

What you graduate with

• 2 GitHub repos with modular, documented code
• Logging, tracing, and monitoring dashboards in both projects
• Automated evals and quality metrics
• 2 live deployed URLs you can share in job applications
• Architecture diagrams + READMEs written for hiring managers
• A demo script and interview narrative for each project

Who this is for

Software, data, or ML engineers with Python experience who want to move into AI engineering roles. You have played with LLMs or built a basic RAG app. You know it is not enough. You are ready to build the real thing.

Program details

• Duration: 6 weeks, weekends only (Sat + Sun, 2.5 hours each)
• Format: Live, online, small cohort
• Stack: Python, LangChain + LangGraph, Google ADK, OpenAI/Anthropic APIs, Langsmith, FastAPI
• Investment: $750
• Seats: Limited

Enroll / ask questions

Chat with us on WhatsApp: Enroll via WhatsApp