Cerebrum
Production LLM training and serving with Mixture-of-Refusals safety routing, FSDP/DDP, vLLM, and Kubernetes deployment.
End-to-end LLM system with distributed FSDP/DDP training, Flash Attention v2, and Mixture-of-Experts support. Includes Mixture-of-Refusals (MoR), a safety routing mechanism achieving 2–3× speedup on safe queries. vLLM inference with quantization, speculative decoding, and prefix caching, deployed on Kubernetes.