🐧 Unify-Agent: An end-to-end unified multimodal agent for faithful, knowledge-grounded image generation.

Structured ecosystem of 190+ AI systems spanning foundation models, agentic reasoning, reinforcement learning, generative architectures, scientific ML, and self-evolving neural systems. A research-driven lab exploring scalable intelligence.

Adaptive parallel reasoning skills for Claude Code.

Reference implementation for "RefTrace: Grounded Normative Reasoning via Retrieval-Augmented Agentic Graph-of-Thought for Multimodal Video Adjudication." Agentic retrieval-loop architecture with Game-State Temporal Hypergraph (GSTH), 6-action typed grammar, Normalized Trace Reward (NTR), and two-stage SFT→GRPO training for NFL penalty adjudication.

Reference implementation for "StateTrace: Explicit State Modeling for Verifiable Multimodal Rule Reasoning." State-bottlenecked reasoning framework with typed state schema, constrained action decoding, transition verification, and dense transition-level rewards for multimodal sports rule adjudication.

Pipeline to investigate structured reasoning and instruction adherence in Vision-Language Models