things I've built

Real products, not demos.

Practical builds across AI agents, finance/life tracking, local-first media tools, VR, and production AI video, focused on shipped systems with clear user value.

Selected work

Chapeta: main product screen 1 / 6

Native macOS AI agent Active product

Chapeta

A full macOS agent harness for computer control: models, tools, screenshots, files, terminal commands, AppleScript, subagents, and approval-based automation behind one shortcut.

  • Built the SwiftUI/AppKit menu-bar UX: hotkey panel, tabs, local history, attachments, speech, themes, and model selection.
  • Implemented the core agent primitives: tool orchestration, file ops, shell, app automation, web/screenshot context, subagents, cancellation, and safety approvals.
  • SwiftUI/AppKit
  • LLM tool orchestration
  • Safety approvals
  • Local-first macOS UX
1 / 5

AI life-tracking SaaS Active build

Gepeta

AI life tracker that turns text, voice, photos, and spreadsheets into structured finance, health, and productivity records users can query.

  • Built the LLM pipeline: router, extractor, tags, image/spreadsheet/bulk flows, and DynamoDB persistence.
  • Shipped React/TanStack PWA + Capacitor iOS with Google OAuth, speech input, push, privacy mode, and analytics dashboards.
  • AI data extraction
  • 13 active life aspects
  • DynamoDB + typed schemas
  • PWA + Capacitor iOS
1 / 4

Privacy-first browser editor Shipped product

FastEdit

Local-first browser editor for converting, compressing, cropping, trimming, blurring, batching, and exporting media without uploads.

  • Built a WebAssembly/browser media pipeline for images, PDFs, animations, and short video.
  • Supports 15 input families, 12 export formats, 79 platform presets, fit-to-size encoding, compare view, and background removal.
  • 100% client-side processing
  • 15 inputs / 12 exports
  • 79 platform presets
  • WebAssembly media pipeline
1 / 4

VR game systems Reality Shift

Terragon: Symbol Of Magic

Unity/OpenXR VR game systems plus an end-to-end CNN gesture pipeline for recognizing user-drawn hand input from raw 3D coordinates.

  • Built the gesture-recognition flow: raw 3D coordinate normalization, dataset preparation, CNN training, and runtime detection.
  • Owned VR input, interactions, combat encounters, procedural levels, enemies, animation timing, and audio sync.
  • Unity + OpenXR
  • CNN gesture recognition
  • 3D coordinate normalization
  • Procedural levels + enemy AI

AI research automation Q.E.D Science

Editorial Grant Agent

End-to-end AI workflow for editorial grant discovery, eligibility screening, evidence gathering, and application drafting.

  • Owned the full pipe: frontend product flow, backend orchestration, data flow, and the AI engine in the middle.
  • Built tool-backed AI workflows for discovery, fit checks, evidence collection, drafting, review, and repeatable research operations.
  • End-to-end product pipe
  • AI engine + tools
  • Eligibility reasoning
  • Draft automation
D-ID: Studio product interface 1 / 3

AI video product engineering Professional work

D-ID Product Foundations

Production engineering across D-ID Studio, Chat, Agents, and API-facing AI video product surfaces.

  • Helped build foundations for Studio creation tools, Chat experiences, and Agents.
  • Led delivery and integration of frontend architecture and API-driven AI workflows.
  • AI video product UX
  • React/TypeScript
  • API integration
  • Product architecture at scale