Lead AI Engineer · delta.exchange
Building AI support and product workflows for a crypto derivatives platform where correctness and responsiveness both matter.
Leading AI at delta.exchange
8 years shipping delightful interfaces and rock-solid systems.
Mostly built to solve my own problems.
Terminal UI that bootstraps a fresh Mac or Linux dev machine — installs your toolchain, restores dotfiles, and distributes SSH identity.
View on GitHub 02macOS menubar app to search and manage Infisical secrets without the dock-icon noise. Native Swift, keyboard-first.
View on GitHub 03A small corner of the internet for gratitude notes — anyone can drop one, no signup, no algorithm. An infinite wall of things people are thankful for.
Visit site 04Six-hour internet experiment for live-music memories. 40+ users on day one with zero ads. Just taste and shipping momentum.
Visit site 05AI portfolio website builder. 87 GitHub stars, 1,000+ users. Built with Anthropic, OAuth, Supabase, deployed on Coolify.
View on GitHubAcross fintech, B2B SaaS, and ecommerce.
Building AI support and product workflows for a crypto derivatives platform where correctness and responsiveness both matter.
Built social listening and influencer workflow tooling, shipped AI features like intelligent naming, sentiment analysis, and address detection, and tightened CI + delivery confidence.
First frontend engineer in the room, later a lead. Scaled the product from MVP to 40+ production features, served 60k+ merchants, built Bot9's conversational AI, and grew the team from 1 to 15+ engineers.
Cut teeth across frontend delivery, mobile apps, and ETL-heavy enterprise work. Good place to learn how to ship under constraints and not panic when requirements move.
Make it useful, then make it beautiful, then make it harder to break.
The frontend is where strategy meets the user. Every layout choice, transition, and empty state quietly decides whether someone trusts the product in the first ten seconds — or never comes back.
AI is interesting when it makes a product feel inevitable, not when it shows off the model. The hard work is restraint: knowing where it adds leverage, and where it just adds latency.
The interesting work lives at the seams — sync, retries, auth, the parts that decide whether the polished frontend actually survives a Tuesday afternoon under real load.
Mentoring, hiring loops, review culture, support SOPs, and helping teams move faster without lowering standards.
If someone needs a walkthrough for the happy path, the product is still lying to itself.
Not just render time. Decision speed, iteration speed, support speed, and how quickly a team can ship the next thing.
And make it feel good while doing it.
I'm up for smart product conversations, sharp design discussions, and ambitious internet ideas that need both engineering depth and taste.
Would you like to read my newspaper? Or just the happy parts?