About
10+
Products / prototypes shipped
7
Agent-driven systems built
500+
Equities screened in finance workflows
5
Schema-aware data connectors
AI Systems Engineer · Full-Stack Builder
|
I build AI products with teeth: agent systems that touch real data, survive messy inputs, and return outputs you can actually trust. Less demo magic, more architecture, evidence, and systems that hold up under load.
Selected projects — 2024–2026
Most AI analytics tools stop at dashboards. Signal Scout spins up a swarm of database-connected investigators that find ICP patterns, test them statistically, and return conclusions with evidence instead of vibes.
Single AI video clips are easy. Coherent films are hard. film-cli gives agents a quality-gated production loop for multi-shot storytelling, continuity control, and shot-by-shot iteration.
A market data app built for glancing, not tapping. Live crypto, stocks, and futures stream straight to the Dynamic Island so you can check price action without opening a brokerage app.
An AI content team for brands that can't afford one: trend research, audience strategy, creative direction, and marketplace-ready copy generated through orchestrated specialist agents.
Walkie-talkies are terrible software. venue-ops-ai turns messy venue operations into a conversational command layer that triages issues, dispatches work, and keeps a live audit trail.
Surveys are rigid and shallow. TokForm makes research feel like a feed: adaptive AI interviews, swipe-native UX, and responses that go materially deeper than static forms.
A negotiation copilot for creators who are tired of getting outpriced. Multi-agent workflows research brands, benchmark deal value, find contacts, and generate smarter counteroffers.
Generic chatbots are boring. skills-chat turns personas, tools, and reusable skills into a modular AI workspace where the assistant can actually do domain-specific work.
Ask a financial question in plain English, get back a reusable analysis artifact. Fundley combines verified data, sandboxed code execution, and agent orchestration in one research workspace.
A calmer take on expense tracking for freelancers: capture receipts, classify deductions, and manage tax-season chaos across web and mobile from one shared system.
About
10+
Products / prototypes shipped
7
Agent-driven systems built
500+
Equities screened in finance workflows
5
Schema-aware data connectors
I build AI systems for domains where bluffing gets expensive: investment research, analytics, operations, and developer workflows. My background is a mix of quantitative finance, backend engineering, and product-minded shipping.
The common thread across these projects is simple: make the model earn its keep. That means structured outputs, sandboxed execution, reproducible pipelines, and interfaces that feel sharp enough to ship — not just cool enough to demo.
Contact