Inspired by Derek Sivers' /now — current, not curated. If anything here looks stale, that just means I've been shipping instead of writing.
Co-Founder & CTO of Griffin AI Tech — ambient intelligence for the physical world. One edge-AI core, three products: InfiniPulse (privacy-first senior-care sensing), InfiniTraq (live retail intelligence from existing cameras), and InfiniDine (real-time dining-floor visibility). Current focus: sub-5-second alerting without ever sending video to the cloud, and expanding deployments across Asia and Europe.
Leading data engineering, mentoring the team, and tightening the Spark-on-EMR pipeline for logistics anomaly + weight prediction.
Taking on a small number of part-time projects alongside Griffin AI — data platforms, applied AI, and technical advisory work.
How far you can push inference quality on a $200 NVIDIA Jetson before having to phone home — and what the right hand-off model looks like for safety-critical alerts.
What does "premium independent" look like when the alternative is a full team and a six-month engagement? Where to draw lines so the side work compounds Griffin AI instead of diluting it.
The next batch of essays is going to be longer and rarer — one signature piece a quarter rather than a flurry of shorter posts.
A mix of operating-manual books and founder essays — leaning toward execution-focused material rather than vision-y reads. (I keep a longer list of all-time favourites on the About page.)