YDLite
I turned yt-dlp and ffmpeg into a focused Windows app, working through startup speed, dependency checks, format compatibility, playlist control, and visible download progress.
AI Product Engineer
I'm Jon Brown. I build local-first tools, creative products, and AI workflows that remain understandable after the demo ends.
Selected proof
YDLite
I turned yt-dlp and ffmpeg into a focused Windows app, working through startup speed, dependency checks, format compatibility, playlist control, and visible download progress.
openResume
I designed a local-first resume workspace around long-term editing: Markdown and block workflows, real-time A4 preview, AI-assisted import, and PDF, DOCX, and HTML export.
BananaCanvas
I built a multimodal workspace where chat and an infinite canvas share one creative workflow, with server-side authentication, durable asset storage, and clear boundaries around generated content.
Capabilities
Turning an early idea into a usable product: interaction design, frontend and backend implementation, data boundaries, deployment, and the failure paths around the happy path.
Designing agent and automation systems with structured inputs, explicit tools, human review, observable failures, and deterministic controls where uncertainty is unacceptable.
Testing emerging tools against real workflows, identifying the useful operating model underneath the hype, and turning the result into clear technical writing.
Toolkit
Working principles
Current focus · July 2026
I am refining local-first product workflows, studying how agent systems fail in practice, and publishing the engineering decisions behind the tools I build.
Core tools: TypeScript, React, Next.js, Tauri, Supabase, PostgreSQL, Docker, Cloudflare, and n8n.
I'm open to focused product engineering work, AI workflow consulting, and open-source collaborations with a clear problem and a practical path to shipping.