About
I like neural networks because they feel less like software and more like small synthetic organisms that learn from data. Most of my work lives close to the model layer: training loops, datasets, loss curves, evaluation, inference, and the strange little behaviours that appear when models start to work. Current interests: - small language models - transformer training - post training - agentic systems - retrieval augmented generation - model evaluation
Links
Badges
Maker History
🎉
Joined Product HuntMarch 4th, 2026
Forums
View more
