Founder of Pixedi. I build production systems with AI and publish the engineering behind them.
Most of my work sits in one gap: the distance between a model that works in a demo and a system that still works on a Tuesday afternoon when nobody is watching it. A job status is the system's opinion of itself, not a fact about the world. A renderer can crash and still report success. An agent can report DONE having written nothing. The fix was never a better prompt. It was a check that reads the world instead of trusting the report.
What I work on
- Browser-based 3D product configurators, and the offline reference renders they get measured against
- Automation infrastructure: verification gates, queues, and pipelines that count what actually shipped
- Applied AI in production, including the parts that did not survive measurement
Open source
- SyncTile for MiniMax H3: 4.9x faster local video generation and seam-free 4K (3840 x 2160) on one RTX 4090, by blending tiles in latent space after every sampling step. Benchmarks against MMH3 Split Upscale and SeedVR2 included.
Write-up: pixedi.com/lab
Writing and measurements: alaintural.com
Most of my work is client work and lives in private repositories. What I publish here is the part that stands on its own: small tools, and the measurement harnesses behind the numbers I write about.