Engineer at heart • Exploring AI, robotics & the human experience
I love building things, exploring, and learning. I write about robotics, AI, spiritual growth and mental health.
This space is where I explore the intersection of technology and humanity, from training neural networks to understanding our own neural pathways.
2025 — Enterprise RAG with 89% accuracy. FastSentiment API with sub-100ms inference. AI research agent processing 100+ papers daily.
2024 — ML systems engineering. Traffic sign classification at 96% accuracy. Attention mechanism visualizer for education.
Physical AI deployment will not be paced by chips, data, or labor. It will be paced by precision reducers, small Japanese-made geared parts most engineers can't name. That reshapes timeline forecasting, industrial policy, and embodied AI safety in ways the discourse hasn't yet caught up to.
AI coding tools have separated into two classes. One compensates for missing operator skill; one amplifies present operator skill. Senior engineers systematically prefer the second. The split is a market signal that operator skill is now the bottleneck, and a snapshot, not a destination.
When the cost of the next step approaches zero, the question of whether to take it stops getting asked. Slow isn't a virtue. It's the protocol for keeping fast aligned with what you actually wanted.