Research Intern - Coding Agents & Post-Training Optimization
Core
Research Intern conducting inquiry and theory into practice, contributing to research and development strides alongside doctoral candidates and researchers.
Role type
Research Intern (LLM Post-Training & Coding Agents)
Builds
Research findings and contributions to the community
Domain
Computer Science / AI / LLMs
Deliverable
research
Required skills
LLM post-training (RL, SFT, preference optimization), model compression, efficient inference, Python, PyTorch, transformer model training/fine-tuning, RL for LLMs at scale, code-generation models, coding agents, quantization, distillation, pruning, speculative decoding, distributed training, GPU profiling, CUDA, Triton
Preferred skills
First-author publications at top venues, experience with serving stacks (vLLM, SGLang, TensorRT-LLM, ONNX Runtime), experience collaborating across teams and disciplines
Technologies
TRL, verl, OpenRLHF, DeepSpeed-Chat, SWE-bench, LiveCodeBench, Terminal-Bench, HumanEval, FSDP, DeepSpeed, Megatron, vLLM, SGLang, TensorRT-LLM, ONNX Runtime, CUDA, Triton
Responsibilities
Collaborate with mentors and other researchers, present findings, contribute to the vibrant life of the community
Seniority
Intern
