Eunice Labs
Exploring the frontiers of synthetic intelligence through curiosity-driven experimentation.
The Lab
Eunice Labs is an independent AI research initiative run by Gagan, focused on understanding how large language models actually work — particularly their reasoning capabilities, agent reliability, and self-modifying systems.
Everything we build is open source. Documenting what works (and what breaks). The goal is simple: push the boundaries of what's possible with LLMs while making the research accessible to everyone.
Current Focus
We are currently focused on reinforcement learning — from policy-gradient fundamentals to how modern reasoning and coding agents are trained.
Alongside that, we are exploring harness engineering and agent memory: how tool-use loops stay reliable over long horizons, and what it actually takes for an agent to remember and build on its own past context. Beyond the hype, the goal is real depth in how these systems learn and operate — not repeating what papers claim.
