Binfeng Xu
Scholar

Binfeng Xu

Google Scholar ID: X8gAsg8AAAAJ
NVIDIA
Machine LearningNLPLLM
Homepage↗Google Scholar↗
Citations & Impact
All-time
Citations
180
 
H-index
4
 
i10-index
4
 
Publications
7
 
Co-authors
0
 
Contact
TwitterOpen ↗GitHubOpen ↗LinkedInOpen ↗
Publications
2 items
ProRL Agent: Rollout-as-a-Service for RL Training of Multi-Turn LLM Agents
2026
Cited
0
Dynamic Noise Preference Optimization for LLM Self-Improvement via Synthetic Data
2025
Cited
0
Resume (English only)
Academic Achievements
  • - Papers:
  • - Gentopia: A Collaborative Platform for Tool-Augmented LLMs
  • - ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Models
  • - Dynamic Noise Preference Optimization for LLM Self-Improvement via Synthetic Data
  • - Efficient Computation of Tucker Decomposition of Correlation-Based Tensors
  • - Iterative Constringency Optimization: Preclustering Approach to Agent Interactive Data
Research Experience
  • - NVIDIA: Working on autonomous agents
  • - Samsung Research America: Built LLM post-training infrastructure and efficient agentic models
  • - New York University: Briefly involved in autonomous driving research
  • - Wake Forest University: Worked on efficient tensor decomposition and object detection for drone and satellite images
Education
  • - Master's Degree: New York University
  • - Advisors: Alfredo Canziani, Yann LeCun
  • - Time: Not specified
  • - Specialization: Autonomous Driving
  • - Bachelor's Degree: Wake Forest University
  • - Advisors: Grey Ballard, Paúl Pauca
  • - Specialization: Efficient Tensor Decomposition, Object Detection for Drone & Satellite Images
Background
  • An AI research engineer. Currently working at NVIDIA on autonomous agents. Formerly a researcher at Samsung Research America (SRA), where he built LLM post-training infrastructure and efficient agentic models. During his MS at NYU, he briefly worked with Alfredo Canziani and Yann LeCun on autonomous driving. At WFU, he was advised by Grey Ballard on efficient tensor decomposition and by Paúl Pauca on object detection for drone & satellite images. In his spare time, he enjoys ML competitions on Kaggle and has ranked in the top 1% globally.
Miscellany
  • Petting Kobu, the Norway Forest cat; Cyberpunk; Digital nomad (someday); Fan of all games by Hidetaka Miyazaki, who motivated him once into indie game development.
Co-authors
0 total
Co-authors: 0 (list not available)