Self-healing code reasoning engine. Detective → QA → Patcher closed loop on SWE-bench, with an RL layer that turns every reasoning trace into DPO training data.
python typescript reinforcement-learning pytest developer-tools autonomous-agents code-intelligence ai-engineering automated-debugging code-repair llm tree-of-thoughts llm-agents agentic-ai ai-coding-assistant swe-bench self-healing-code dpo-training reasoning-traces closed-loop-reasoning
-
Updated
Aug 21, 2026 - TypeScript