Scholar
Zhilin Wang
Google Scholar ID: OmMgSQsAAAAJ
Senior Applied Scientist, NVIDIA
Citations & Impact
Citations
828
H-index
12
i10-index
12
Publications
20
Co-authors
5
list available
Publications
Browse publications on Google Scholar (top-right) ↗
Resume (English only)
Academic Achievements
- - Built HelpSteer2-Preference & Llama-3.1-Nemotron-70B models with over 1M downloads (ICLR 2025)
- - Architected Humanoid Agents Simulation Platform (EMNLP 2023 System Demo)
- - Started an EdTech startup, learnAh.org, helping over 500k Chemistry students
- - Top 20 Global Finalist at Google Science Fair with a project on Zinc-Air Batteries
- - Sliver Medalist at Singapore Chemistry Olympiad (SChO) and invited for IChO training team
Research Experience
- - Working on LLM alignment at NVIDIA
- - Leading the development of the HelpSteer dataset series
- - Exploring new data modeling methods, such as creating the Llama-3.1-Nemotron-70B-Instruct and Reward models
- - Developed the Humanoid Agents Simulation Platform
Background
- Senior Applied Scientist at NVIDIA, focusing on aligning LLMs to be more helpful for humans. Recent work focuses on leading the HelpSteer dataset series, which are popular open-source post-training datasets for helpfulness modeling. Additionally, he explores creative ways to model such data, giving rise to the Llama-3.1-Nemotron-70B-Instruct and Reward models, as well as the Llama-3.3-Nemotron-70B model family. Personal hobby projects include Humanoid Agents (a platform for simulating human-like agents).
Miscellany
- Hobby projects: Humanoid Agents (a simulation platform for human-like agents)
