Tuned Tensor
Train and fine tune small local models for focused tasks. One spec file drives training, evaluation and serving on your own GPU.
Work
Products and experiments.
Three products I’m building, plus smaller experiments.
Products
Hedge Layer
Financial infrastructure with AI built in. Agent H researches prediction markets and replies with live prices, cited evidence and a probability range you can check.
Deeper Level
Play strategy games against AI models. Each game becomes reinforcement learning data for training better local models.
Experiments
Autoresearch Classifier
An automated experiment loop that searches for better prompt injection classifiers using scikit-learn.
Bedrock Agent Blueprint
A repeatable way to deploy AI agents on AWS, with infrastructure defined in code.
Mental Model Graph
Explore how mental models from economics, psychology, mathematics and logic connect.