AI Agent Factory
AI Agent Factory: Build Autonomous AI Systems with an AI Startup. A 4–8 weeks Intravarsity career experience (30–50 hours) for school and undergraduate studen
You'll work alongside a real AI startup building production AI systems inside our San Francisco campus. Using the same APIs and frameworks they use daily — OpenAI, Anthropic, LangChain — you'll construct AI agents that autonomously complete complex tasks. Your capstone project solves a real problem for the startup, evaluated by their CTO.
Go beyond ChatGPT. Embedded with a real AI startup on our San Francisco campus AI floor, you'll build autonomous agents that reason, plan, and use tools — the same systems powering the next generation of products.
Who it is for
High school, secondary school students & undergraduates. Basic Python required. Passion for AI essential.
Weekly plan
Week 1: LLM Foundations & Prompt Engineering
- Tour the AI startup's office and understand their production AI stack
- Master advanced prompt engineering — chain-of-thought, few-shot, function calling
- Build a conversational agent with memory using Python
- Implement structured output parsing for real-world data extraction
Week 2: Tool-Using & RAG Agents
- Give your AI agent access to web search, databases, and calculators
- Build a Retrieval Augmented Generation system with vector embeddings
- Implement the ReAct framework for autonomous research tasks
- Connect your agent to a startup's actual knowledge base
Week 3: Multi-Agent Systems
- Design multi-agent architectures where AI agents collaborate
- Build a system where agents delegate, debate, and converge on answers
- Work on a real startup problem — automating a workflow with AI agents
- Optimize performance, cost, and latency with the engineering team
Week 4: Ship & Demo
- Build your capstone: an AI agent solving a real startup problem
- Test edge cases with feedback from the startup's engineers
- Present to AI researchers, startup CTOs, and venture partners
- Receive startup recommendation and Proof-of-Work credential
What you produce
- Working AI agent system solving a real problem
- Technical documentation and architecture diagram
- Demo video
- AI startup recommendation letter
- Intravarsity Proof-of-Work credential
Skills
Python, Prompt Engineering, LangChain, RAG Systems, Vector Databases, Agent Architecture, API Integration, Production AI
Commitment
~5–8 hours/week for 4–8 weeks (30–50 hours total). In-person in San Francisco with the host company.