Backend systems and AI agents. Ship before polish.
Each tile below is a live preview of something I actually shipped — animation states model the real behaviour. Click through for the post-mortems.
AI-powered platform that transforms any source content—blog, PDF, docx, URL—into platform-optimized variations in under 10 seconds.
Scan barcodes or label text for instant, science-backed health ratings—making smarter choices effortless.
Autonomous corporate integrity auditor—scrapes SEC filings, cross-references live web signals, and calculates a Decoupling Coefficient with screenshot proof.
Tools are arguments. These are the ones I'm willing to defend in a code review at 11pm. Substitutes welcome — the ratio of opinion to dogma matters more than the brand on the lid.
What I reach for first.
Where the AI work lives.
How it gets to a user this week.
Years employed is a useless metric. Things shipped is the metric. Recent first.
Shipped Content Repurposing Engine. AI-powered content transformation platform.
Built Health Label. Instant science-backed health ratings from barcode scans.
Created Alpha Audit. Autonomous corporate integrity auditor with SEC cross-referencing.
Looking for the next thing to break in production for the right reason.
Most AI work in 2026 is still a demo dressed as a product. The interesting work is the part nobody posts about: the retry logic, the idempotency, the cache that survives a bad deploy. I want to be the engineer you call when the wrapper breaks and there's nothing on Stack Overflow yet.