Topic: How I Run My Life (and Business) on AI Agents
Description:
Jiquan Ngiam, founder and CEO of MintMCP, shares how he uses AI agents to manage both personal and professional workflows. Drawing from his experience at Google Brain, Stanford, and as an early contributor to Coursera, Jiquan focuses on the practical side of agentic AI rather than research prototypes. He introduces a simple recipe for building effective agents (combining a runtime harness, persistent memory, and secure data access) and demonstrates real-world use cases such as tracking health metrics, managing to-do lists, remembering personal and professional contacts, and keeping business processes moving forward. The session also covers concrete implementation patterns, including instruction files (e.g., CLAUDE.md), using Git as agent memory, and Model Context Protocols (MCPs) for connecting agents to external tools and data, along with an honest discussion of security risks when agents handle sensitive information.
Recording
Key Takeaways
A practical framework for building useful AI agents using a runtime harness, persistent memory, and data access.
Implementation patterns you can use today, including instruction files, Git-based memory, and MCPs for tool and data integration.
Real-world examples of agents managing health, tasks, relationships, and business workflows—plus key security considerations when giving agents access to sensitive data.
Additional Info
Here is a link to Jiquan's Life System Setup Github Repo:
https://github.com/jngiam/life-system-setup
