Low Precision RL
The unified FP8 path: matched quantization between training and inference,
BF16 backward and master weights.
INT4 QAT
W4A16 quantization-aware training for fitting large models on a single
8-GPU node.
Rollout Routing Replay (R3)
Capture expert routing during inference and replay during training. The
mechanism that keeps MoE RL stable.
Speculative Decoding
Draft + target speculative rollout, with online MTP-SFT for the draft.
On-Policy Distillation
Train a student on its own rollouts while matching teacher token
probabilities through SGLang or Megatron teacher modes.
LoRA Training and Serving
Train LoRA adapters with SFT or RL and serve them through SGLang from the
same checkpoint.