Full stack & AI engineer · NYC

Full stack apps, now with agents inside

I'm Song. I build full stack web apps and LLM-powered agents — React frontends, Node APIs, and Claude agents with tool use and memory. Open to full stack and AI engineering roles.

verified
Open to work
NYC · hybrid or remote
code

Full stack web apps

React and TypeScript on the front, Node and Express APIs behind them.

smart_toy

Agents with real tool use

Claude agents that call tools, validate input and persist what they collect.

storage

APIs and data models

REST and GraphQL endpoints over MongoDB and Supabase schemas.

rocket_launch

Shipped, not sketched

Three products taken from a rough idea to a live URL.

What I build

One engineer across the whole slice

Schema, API, interface and the agent loop between them — shipped as working products rather than demos.

React TypeScript Node.js Python Claude API MongoDB Supabase GraphQL

Selected work

Things I've shipped end to end

Pitch Picture screenshot Pitch Picture screenshot

01

Pitch Picture

Turns a spoken brainstorm into a visual diagram — Deepgram transcribes, Claude picks a diagram type and writes Mermaid, the app renders it.

Claude AI Deepgram React 19 Express Supabase
Autonomous Agent Reviewers screenshot Autonomous Agent Reviewers screenshot

02

Autonomous Agent Reviewers

A swarm of AI judge agents that peer-review hackathon submissions across two rounds of deliberation, with every verdict recorded on 0G Storage as an auditable artifact. Three specialized judges — technical, originality, skeptic — revise their assessments after seeing their peers' evaluations.

Claude AI React TypeScript Node.js Express 0G Storage
Customer Onboard Agent screenshot Customer Onboard Agent screenshot

03

Customer Onboard Agent

Replaces onboarding forms with a conversation. Claude tool use extracts, validates and saves customer fields in real time from a configurable schema.

Claude AI React TypeScript Supabase Edge Functions
send

Let's build something

Looking for full stack or AI engineering work — startups building with LLMs, or teams that move fast and care about the craft.