- Released SFR-DeepResearch, LiveReserachBench, Hard2Verify, Variation in Verification
- Ruochen Zhao and Chengwei Qin have recently graduated and joined Apple (Research Engineer) and HKUST (Asst. Prof), respectively. Congrats!
- Received a presidential award FY'25 from Salesforce (one of the 0.5% employees!)
- Released Process DPO from Pseudo Feedback, Fin-DAP, SFR-RAG, SFR-Judge, GemFilter, FaithEval, and CRM Benchmarks
- Gave in-person invited talk at JHU CSLP
- Gave a tutorial on LLM+Vis at EMNLP-23 and four keynotes on LLMs: (a) LLM-AI workshop @KDD-23, (b) KAUST Research Conference, (c) QCRI LLM Seminar, and (d) Taming LLMs workshop @SIGDIAL-23 (slides)
- Released XGen-7B LLM that supports 8K sequence length (Code, Blog). Good at both text and code, and specially on long sequence tasks
Research Experience
- Senior Research Director, Salesforce AI Research, Palo Alto, USA, [Feb'25 - ]
- Associate Professor (tenured, on leave since Aug 2022), Nanyang Technological University (NTU), Singapore [Mar'22 - ]
- Research Director, Salesforce AI Research, Palo Alto, USA [Sep'22 - Jan'25]
- Senior Research Manager (part time), Salesforce AI Research, Singapore [Feb'19 - Aug'22]
- Assistant Professor, Nanyang Technological University (NTU), Singapore [Jul'17 - Feb'22]
- Research Scientist, Qatar Computing Research Institute (QCRI), Doha [Jan'14 - Jul'17]
Education
- PhD in Computer Science, University of British Columbia (UBC), Sep'08 - Feb'14
Background
- Research Interests: Natural Language Processing, Large Language Models, Multi-lingual NLP (Machine Translation, Cross-lingual tasks), Multi-modal NLP (NLP+Code, NLP+Vision), Robust/adversarial NLP, Machine Learning, Deep Learning, Reinforcement Learning
- Professional Field: Computer Science
- Brief Introduction: Senior Research Director at Salesforce AI Research, previously an Associate Professor at Nanyang Technological University, currently on leave.
Miscellany
- Created website with git, jekyll, bootstrap, and sublime text