About
I'm a Senior Machine Learning Engineer at Meta, leading development of Facebook's Feed ranking systems and LLM-driven content discovery for 1.8B+ monthly active users. My work spans the full ranking stack — retrieval, early ranking, and value models — and increasingly explores agent-centric architectures that bring LLM reasoning to recommendation.
Before Meta, I earned my Ph.D. in computer vision at Indiana University, advised by Prof. David Crandall, focusing on egocentric (first-person) video understanding and cross-view perception. I contributed to large-scale initiatives including Ego4D, with work spanning activity recognition, person identification, autonomous systems, and remote sensing. My research has appeared at CVPR, ECCV, and IEEE TPAMI and has drawn over 3,000 citations.
Experience
Senior Machine Learning Engineer — Recommendations & Ranking
2024 – PresentMeta
- Lead end-to-end development of Facebook's connected and unconnected Feed ranking systems and LLM-driven content discovery, managing large-scale recommendation engines serving 1.8B+ monthly active users.
- Pioneered an agent-centric architecture that brings LLM-based reasoning to ranking pipelines for natural-language preference processing — an agentic recommendation system delivering +20% NDCG@10 over MemRec and +108% over memory-less LLMs.
- Developed content-quality classifiers to detect inauthentic and coordinated engagement, strengthening platform integrity and securing ~$5M in annualized payout savings.
- Own Feed ranking optimization across retrieval, early ranking, and Pass-1 value models; define the boost/demotion experimentation roadmap and run 100M-scale A/B tests, driving lifts in DAU, revenue, sessions, and time spent.
- Previously interned at Meta (2022), improving sourcing and ranking for the Instagram Reels creation-trends surface (+10% normalized entropy).
Self-Driving Researcher Intern
2021Baidu
- Analyzed failed lane-change cases across NGSIM, HighD, Waymo, and internal logs, and developed lane-change planning and prediction algorithms that outperformed the prior rule-based model.
Publications
-
Fusing Personal and Environmental Cues for Identification and Segmentation of First-Person Camera Wearers in Third-Person Views
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024
-
Ego4D: Around the World in 3,000 Hours of Egocentric Video
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022 Oral paper
-
DoTA: Unsupervised Detection of Traffic Anomaly in Driving Videos
IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2022 paper
-
Stepwise Goal-Driven Networks for Trajectory Prediction
IEEE Robotics and Automation Letters (RA-L), 2022, with ICRA 2022. 3rd on nuScenes prediction task, 6th AI Driving Olympics, ICRA 2021. paper
-
Deep Tiered Image Segmentation for Detecting Internal Ice Layers in Radar Imagery
IEEE International Conference on Multimedia and Expo (ICME), 2021 Oral paper
-
Unsupervised Traffic Accident Detection in First-Person Videos
International Conference on Intelligent Robots and Systems (IROS), 2019 paper
-
Joint Person Segmentation and Identification in Synchronized First- and Third-person Videos
European Conference on Computer Vision (ECCV), 2018 paper
Education
-
2017 – 2024
Ph.D. in Intelligent Systems Engineering
Indiana University Bloomington · Minor in Statistical Science
-
2015 – 2017
M.S. in Computer Science
Indiana University Bloomington
-
2007 – 2011
B.S. in Software Engineering
Tongji University
Teaching
Associate Instructor — Indiana University
- FA16 CSCI B561 — Advanced Database Concepts
- SP17 CSCI P536 — Advanced Operating Systems
- FA17 ISE E101 — Innovation and Design
- SP18 ISE E599 — Topics in ISE (Advanced Operating Systems)
- FA18 ISE E599 — Topics in ISE (Advanced Operating Systems)