Projects
Mailman
WIP
A deep dive into RL on LLMs. Three environments, data generation, synthetic databases, RL trainers, and more.
terminal2F
WIP
A research platform and lab for agents and agent systems.
OUDAU
The startup that never started
The allocation layer for AI resources. [compute markets]
neural-geometry
Building intuition for neural networks, with simple neural networks, visualization and graphics programming.
P2Engine
Thesis
A framework + runtime to build, run, and evaluate multi-agent systems. Extended with the Canton Network to enable monetary incentives, payments, and audits. The artifact of my master's thesis.
