Hey, I'm an applied RL / 3d engineer working on simulations. I'm currently a research scientist at the MIT Media Lab cs sf lab and getting a Masters in Data Science at Harvard '27 exp.
For a recent project, try out text-to-3d-simulation on a map against real-time LiveUA Map data.
I also migrated Unity's ML-Agents to Three.js+Python. These run PPO mostly using SB3 and train other various RL and SSL models in Python and Torch and then use Three.js as a visualization layer.
Recently, I've been using RL+SSL for training autonomous agents in simulations to study causal inference across diverse domains such as the SF 20k Personas Dataset. Email me if interested for more on this.