Zengxiao (Henry) He

Building vertical AI SaaS · Stanford EECS · Agentic Workflows


About

I’m a builder and engineer currently exploring the intersection of vertical SaaS and autonomous AI agents. I’m building a stealth startup focused on engineering robust multi-agent systems that solve complex, real-world industry bottlenecks.

My research and engineering focus centers on agentic workflows, specifically, designing scalable, stateful agents that transition AI from conversational novelty to operational automation.

When I’m not coding or architecting systems, you can find me playing fingerstyle guitar, on the tennis court, or pursuing my goal of reaching race-car driver levels of performance driving. And if you still can’t find me? I’m probably at the gym (exclusively hitting upper body push/pull days; we don’t talk about leg day).


Experience

Co-Founder

Stealth AI Startup

Jan 2026 — Present

Building vertically integrated AI tools to automate high-friction workflows.

Teaching Assistant (CS 295: Software Engineering)

Stanford University

Dec 2025 — Present

Teaching and mentoring students on building robust software systems alongside Prof. Sara Achour. Designing assignment workflows, reviewing LLM-generated code for correctness, and providing project feedback to students.

Software Engineer Intern — AI/ML

Oracle

Jun 2025 — Sep 2025

Architected an autonomous agent to streamline and execute backport operations within Oracle SCM, assisting engineers by continuously monitoring the BugDB system and automatically running scripts to improve operational throughput by 40%

Research Assistant

Monash University

Aug 2022 — Aug 2023

Researched and implemented transformer-based computer vision models under the guidance of Prof. Jia Wu to assist doctors in the automated diagnosis of osteosarcoma, significantly improving clinical workflow efficiency.


Projects

NotionAgent Chrome Extension

A browser extension that seamlessly integrates localized AI agents into Notion, turning static documentation into executable task workflows.

Auto-Reviewer (CS 295 Tooling)

An automated LLM-based code review pipeline designed to evaluate student software architecture and provide real-time, context-aware feedback.


Education

Stanford University

M.S. in Electrical Engineering

Expected Jun 2026

Central South University

B.S. in Computer Software Engineering

2020 — 2024

Core Member, National Innovation & Entrepreneurship Base (Mentored 10+ startup teams & robotics competition teams). Varsity Member, Software Engineering Basketball Team.