██████╗ █████╗ ███████╗ █████╗ ███╗ ██╗████████╗
██╔══██╗██╔══██╗██╔════╝██╔══██╗████╗ ██║╚══██╔══╝
██████╔╝███████║███████╗███████║██╔██╗ ██║ ██║
██╔══██╗██╔══██║╚════██║██╔══██║██║╚██╗██║ ██║
██████╔╝██║ ██║███████║██║ ██║██║ ╚████║ ██║
╚═════╝ ╚═╝ ╚═╝╚══════╝╚═╝ ╚═╝╚═╝ ╚═══╝ ╚═╝
Full-stack engineer with a product-first mindset. I gravitate toward problems where correctness matters — distributed systems, type-safe APIs, async pipelines that don't silently drop data.
Currently building logistics infrastructure at supertruck.ai — real-time trip management, document parsing pipelines, and carrier integrations that hold up under load.
When I'm not writing code, I'm usually explaining it. I do technical mentorship and enjoy the kind of code reviews where both sides learn something. I prefer a terminal over a GUI, Zod over runtime surprises, and PostgreSQL over everything else.
|
Primary Languages |
Backend Frameworks |
Data Layer |
|
supertruck.ai · production Real-time trip management for trucking ops. Email attachment parsing, rate confirmation workflows, async event handling, carrier integrations. Built the NestJS backend from zero to production.
|
Full-stack · shipped Multi-platform post scheduling dashboard. FastAPI async backend, OAuth account flows, per-platform media specs, presigned R2 uploads. Fought React Strict Mode double-submission so you don't have to.
|
|
High-concurrency · OSS Custom layer on Node.js + Redis Streams. 5,000+ events/sec. Consumer group ACK, dead letter handling, fan-out patterns. Built because off-the-shelf wasn't cutting it.
|
Data viz · shipped Complex analytics surface with D3.js + Prisma + real-time PostgreSQL streaming. The kind of dashboard where the loading state matters as much as the data.
|
Database indexing is a first-class design decision. Most devs add indexes reactively when queries slow down. I treat index design like schema design — upfront, intentional, and documented.
Type-safety is load-bearing, not ceremonial. Zod + Prisma + TypeScript end-to-end isn't over-engineering. It's buying runtime confidence in exchange for compile-time discipline. A trade I'll always take.
The hardest bugs only appear at 3× traffic. Race conditions, unacked queue messages, connection pool exhaustion. Production teaches you what local dev never will. I keep a list.
Good dotfiles beat any GUI tool. My terminal setup is a system. Every alias, function, and prompt line is deliberate. A well-tuned shell over any drag-and-drop interface, every time.
I'm reachable for:
- System design discussions — especially distributed systems, API contracts, and anything where the wrong call at design time costs weeks at runtime
- Technical mentorship — async code review or architecture pairing; I give direct feedback and I remember what it felt like not to know things
- Open source collaboration — NestJS, FastAPI, real-time infra; reach out with a specific problem, not a generic "collab?"
The best opener is a specific question. "I'm building X and I'm stuck on Y" → always gets a response.



