Iacon operates Optimus, the first autonomous RL engineer.

We believe superintelligence will not be a single monolithic model, but a network of domain-specific models—from LLMs to World models to RL policies. The fundamental training mechanism for all of them is: if an agent can practice, observe, and improve within an environment, it can be trained. We are automating this universal loop.

Iacon lets teams set the direction: defining environments, rewards, and success criteria. Optimus, an autonomous RL engineer, takes that objective and runs the full loop directly inside your stack. He conducts rigorous scientific experimentation—implementing sota research or synthesizing entirely novel methods to filter high-signal trajectories from raw environment interactions to train generalizable policies.

Give Optimus a goal, and it delivers a trained agent—works across LLMs, harnesses, diffusion models, robotics, NPCs, biological systems, or custom policies.