rollout-infra

Validate GRPO training configs against 36 cross-framework bug avoidance rules (19 MUST DO + 17 MUST NOT). Covers 7 frameworks (DeepSpeed, Megatron, vLLM, verl, SGLang, rLLM, MindIE), 7 pattern classes, and MoE NaN avoidance.

Debug and diagnose GRPO training issues across verl, Megatron, DeepSpeed, and RTX 4090. Symptom-to-bug mapping, NaN debugging, Muon clipping avoidance, and CUDA stream safety analysis.

Navigate SGLang source code. Quick reference for subsystem locations, key classes, and data flow.

Navigate mechanistic interpretability concepts, tools, and SAE architectures. Quick reference for SAE variants, feature steering methods, circuit discovery, and trust calibration.

Navigate model merging strategies and theory. Quick reference for Task Arithmetic, DARE, TIES merging methods, alpha selection, and RTX 4090 validation results.

Navigate RDMA, InfiniBand, RoCE v2, AI cluster networking topology, and NCCL transport internals. Quick reference for Verbs API, QP/CQ/SRQ/MR, fat-tree/rail-aligned topology, DCQCN congestion control, GPUDirect RDMA, and cluster configuration.

Navigate PyTorch torch.distributed, NCCL, and distributed training internals. Quick reference for ProcessGroupNCCL, DDP, FSDP, NCCL channels/proxy/protocol, and training configuration.

Run GPU experiments on matpool.com A16 server. Pushes script to GPU, runs it, pulls results, creates experiment notes.

Estimate LLM inference latency, throughput, and memory requirements. Roofline model, batch sizing, KV cache analysis.

Navigate Megatron-LM source code. Quick reference for subsystem locations, key classes, and data flow across TP/PP/SP/Communication/Optimizer/Checkpoint/DataLoader/MoE modules.

Navigate vLLM V1 source code. Quick reference for subsystem locations, key classes, and data flow.