LLM Algorithm Expert
Alibaba DAMO Academy
- Location
- Hong Kong
- Level
- Expert
- Posted
- 25 Sep 2026
LLM algorithmExpertOn-siteActive
Original posting
Skills
- LLM
- Post-training
- SFT
- Reasoning
- PyTorch
- Distributed training
Summary
Alibaba DAMO Academy's Hong Kong group is hiring an algorithm expert for post-training of Tongyi-class language models. The charter is reasoning quality: supervised traces, preference data, and rejection sampling, not a new pretraining run from scratch. You will own one post-training stage, set the eval suite before changing the recipe, and review data that vendors deliver. The seat is in Hong Kong, with training jobs launched on the Hangzhou cluster. Requirements: shipped or published work on LLM post-training, strong PyTorch, and a habit of keeping ablations small enough to trust.
Full description
Alibaba DAMO Academy's Hong Kong group is hiring an algorithm expert for post-training of Tongyi-class language models. The charter is reasoning quality: supervised traces, preference data, and rejection sampling, not a new pretraining run from scratch.
You will own one post-training stage, set the eval suite before changing the recipe, and review data that vendors deliver. The seat is in Hong Kong, with training jobs launched on the Hangzhou cluster.
Requirements: shipped or published work on LLM post-training, strong PyTorch, and a habit of keeping ablations small enough to trust.