EdotEnv
Quant Neolab working toward RSI.
EdotEnv's mission is to teach models recursive self-improvement. To get there, we build RL environments at scale from financial market data that increase in difficulty, so agents can continuously hillclimb.
A frontier that keeps moving.
Recursive self-improvement is a loop: an agent researches, learns from the outcome, and returns to the next problem with better tools and judgment. The loop only continues when each round gets harder.
Markets create that pressure naturally. Every useful discovery attracts competition, and every successful strategy makes the next opportunity harder to find. Markets are a natural proving ground for agents that need to keep improving.
RL environments for harder problems.
We capture this dynamic in multi-step RL environments where agents form hypotheses, design experiments, verify results, and iterate. We use them to evaluate and post-train agents on increasingly difficult research tasks.
We work with frontier AI labs and academic groups building research harnesses, evaluation benchmarks, and post-training environments.
Progress should belong to everyone.
We believe recursive self-improvement can help society reach scientific and technological breakthroughs faster—and make the benefits of that progress available regardless of socioeconomic status.
Our long-term goal is to redistribute the gains from faster progress equally to everyone.
Working on RSI, RL, post-training, or AI for research?
We’d like to hear from you.
Join the conversation