Kevin Anderson — Staff Software Engineer building open-source tools for self-improving AI agents and the post-training of open-weight code models. Work spans agent orchestration, reinforcement learning (RLHF/RLVR/GRPO), reward modeling, and agent evaluation.