I build AI tools from real workflow pain, then publish the proof.

I combine product thinking and engineering execution: finding repeatable workflow pain, turning it into AI-native tools, and making the result legible through code, docs, and writing.

Current proof: production QA ownership, a first-author IEEE paper, and public projects across digital twins, agent evaluation, model infrastructure, and source-code research.

P0/P1 incidents across owned production surfaces

1,884

Claude Code source files recovered and analyzed

87%

parameter reduction in my first-author IEEE paper

Walk the proof demo Open proof chain Scan recent ships

Proof-first site

Start from the proof chain, then read the biography.

A visitor should be able to inspect what I have actually built: working repos, demo routes, technical writing, and explicit privacy boundaries. The career timeline is still available, but the homepage now routes through evidence first.

Digital Twin

Inspect →

A file-first operating layer that turns my knowledge base into a working digital twin for research, writing, site updates, and continuous refinement.

Dedicated on-site explainer pageWiki-centered working memory

Agent Scorecard

Inspect →

A trace-first standard and CLI for judging whether an AI agent is worth more token budget, permissions, and autonomy in my real workflows.

Scores evidence, not vibesChecks tools, artifacts, verification, and busywork

CLIProxyAPI

Inspect →

A proxy server that exposes OpenAI/Gemini/Claude/Codex-compatible APIs for CLI and coding tools, with OAuth login, provider routing, and multi-account load balancing.

Codex + Claude Code OAuth supportProvider routing for coding tools

Personal Website

Inspect →

My public narrative system: a Next.js site, MDX writing archive, and AI-assisted publishing workflow designed to compound proof over time.

Static and dynamic deployment pathsMDX blog with search, tags, RSS, and structured data

Open full proof chain Then view experience

Background

About Steven
Builder profile after the proof chain
Experience & Projects
Full-time roles and side projects
Internships
6 internship roles across industries
Education
Academic journey and achievements

Building

Open-source agent proof
Public repos, a public PR, and local/public boundaries
Future
Goals, current focus, and milestone journal

Writing

Latest: Three-Agent Daily Log — 2026-05-14
Product Owner / Builder / User-side Reviewer daily log: proof-chain shipped page, system map, contributions page, scoring dimensions demo, and GitHub profile automation.
Blog archive
Notes on AI tooling, engineering, investing systems, and what I'm learning