AI Engineer at AI Tech Cloud Network (formerly Solidus AI Tech). I work across agent orchestration, developer tools, and data infrastructure, usually end to end, from backend services and APIs to the interfaces people actually use.
// automation · blockchain · developer tools
From AI workflow platforms to Farcaster indexers and mobile games, most of my projects are built end to end — from infrastructure and data pipelines to the UI.
A visual workflow builder for AI automation. Drag blocks onto a canvas, connect models, tools and Human-in-the-loop Blocks and deploy the result as an API, webhook, or scheduled job. I architected the backend and built the meta-agent that converts plain-language instructions into fully functioning workflows.
A guide for Hollow Knight and Silksong that knows where you are in the game and refuses to spoil what's ahead. Mark your progress and ask anything — spoiler control is enforced as vector-store payload filters, not prompts, over 5,000+ wiki chunks with hybrid retrieval (dense + BM25, RRF-fused) and cross-encoder reranking. Evaluated at 100% recall@6 and 0% spoiler-leak, with an in-character guide, two guidance modes, and bring-your-own-key model support.
A high-throughput Farcaster indexer built on a sharded event pipeline and Kafka. I developed the developer APIs and a live Scanner dashboard that streams network activity in real time.
A fast PvP trivia game for iOS, built in Swift — real-time duels, daily topics, ranks and leaderboards. It started as a Farcaster mini-app; it's now a fully standalone native app with a brand-new design.
Hides the blockchain so people can connect, chat and send funds using readable Farcaster or Base names instead of wallet addresses.
A project-funding forum where people submit Marginal Utility Curves for per-dollar impact. The backend is serverless — GitHub Gists linked through Farcaster casts, no database.
I like building the less glamorous parts of AI: the orchestration layers, tool connections, data pipelines, and interfaces that turn impressive demos into products people can actually use.
Before AI, I spent two years solving integration and automation problems at Deloitte. That experience shaped how I approach engineering today: start with the workflow, build reliable systems underneath, and make the complexity disappear for the user.
// off the clock: gaming, marathons, and a lot of reading