TalentFlow
Co-founder · ShippingAn entertainment SaaS I co-founded in November 2025. We are building the operational backbone for talent teams — the scheduling, routing, and payment plumbing that currently lives in spreadsheets and group chats.
I build software. Right now that means co-founding TalentFlow, getting Health Harmony ready to launch, and writing the tooling behind Dorian Trader. This page is the index — what I'm working on, what I'm into, and where to find me.
An entertainment SaaS I co-founded in November 2025. We are building the operational backbone for talent teams — the scheduling, routing, and payment plumbing that currently lives in spreadsheets and group chats.
A personal health management platform getting ready to launch. It pulls the records, labs, and appointments scattered across every portal you have ever been given a login for into one place you actually control.
I am a member of technical staff at Dorian Trader, where I build the tooling behind the community — the bots people talk to, the signals they act on, and the interfaces that connect it all to their brokerage.
The bots members interact with day to day: alerts, lookups, and the automation that keeps the community running without a human in the loop.
The pipeline that turns market data into the signals members act on, and delivers them where people already are.
A Model Context Protocol server for tastytrade, so an LLM agent can read positions and reason about a portfolio through a typed tool interface instead of scraping a UI.
I went to HOPE 17 (Hackers On Planet Earth) this year. A few talks are still rattling around in my head:
My favorite of the weekend — cognitive security as the human-side twin of cybersecurity, and why our mental defenses need the same seriousness we give silicon.
Real onion circuits from a normal browser — Tor for people who cannot install privileged software, and the new risks that comes with.
A free offline Android IDE that turns a budget phone into a full workstation — coding without the laptop gatekeeping.
These are referral links — if you sign up through one, I may get credit for it.
Metered inference looks like a healthy business. Flat-rate plans are where the money is being lost.
Anthropic reviewed 141,006 eval runs to find three incidents. The most useful details are the ones the coverage skipped.
The most AI-generated place on the internet is where everyone goes to prove they understand AI.
The gap between dispatching a task and getting it back is ninety seconds. That interval is where attention goes to die.
One compromised account put a credential stealer into 400+ npm packages. The technical fixes do not touch the cause.