ICLR 2022, 2020
TL;DR: Demonstrates gradient inversion label leakage in vertical split learning and designs the Marvell noise defense mechanism to preserve privacy without degrading utility.
Jiankai Sun
Machine Learning Engineer at Pinterest · Ph.D. in Computer Science (The Ohio State University)
I am a Machine Learning Engineer at Pinterest. Previously, I was a Research Scientist at ByteDance AML-US in Seattle. I obtained my Ph.D. in Computer Science from The Ohio State University, advised by Prof. Srinivasan Parthasarathy.
My research and engineering focus centers on Privacy-Preserving Machine Learning (Federated Learning, Split Learning, Data Clean Rooms), Large-Scale Recommender Systems (Ads ranking, decoupled representation learning, retrieval), Graph Representation Learning, and Large Language Models & Alignment (code generation, proxy fairness). Outside of work, I love vibe coding indie mobile apps, traditional board games, and local-first utilities.
1,639
Citations
20
h-index
23
i10-index
50+
Publications & Patents
Core Research Themes
Scientific and algorithmic focus areas. Click any theme to explore corresponding publications.
01 / PRIVACY
12 papers
Federated & Split Learning
Two-party split learning label protection, cross-silo differential privacy, data clean rooms, and vertical federated learning.
View papers →
02 / RECSYS
7 papers
Recommender Systems
Decoupled representation learning for ads ranking, end-to-end retrieval, machine unlearning, and diversified item ranking.
View papers →
03 / GRAPHS
8 papers
Graph Representation
Directed graph embedding with asymmetric transitivity preservation, cycle breaking in noisy hierarchies, and cold-start routing.
View papers →
04 / LLMS
3 papers
LLMs & AI Alignment
Improving LLM code generation through print debugging, proxy fairness with missing sensitive attributes, and self-refinement.
View papers →
Vibe Coding & Indie Software
Handcrafted native iOS apps, board game engines, and local-first personal utilities.
FolkBoardGames
Game Suite
Traditional folk board game suite (Ou-Zi Qi, Mi-Zi Qi, Xia-Shi Qi, Fang Qi). Zero-latency canvas rendering, offline AI engine, and interactive rules.
TinyTasteBook
iOS App
Baby solids & allergy tracker designed with Swift & SwiftUI. Local-first architecture, zero analytics tracking, clear pediatric timeline, and clean reactions journal.
HoiBaby
iOS Utility
Infant feeding routine tracker with iOS 17 Interactive Widgets and Live Activities. One-tap nursing timers, formula logs, and daily rhythm insights.
Publications Stream
50 works · 32 papers · 18 patents
WSDM 2023
★ HIGHLIGHT
Dgrec: Graph neural network for recommendation with diversified embedding generation
217 citesProceedings of the sixteenth ACM international conference on web search and …, 2023
TL;DR: Diversified graph embedding generation for mitigating filter bubbles in user-item recommendation graphs.
Proceedings of the 26th international conference on World Wide Web, 183-192, 2017
Proceedings of the 2018 SIAM international conference on data mining, 153-161, 2018
CIKM 2021
★ HIGHLIGHT
Learning an end-to-end structure for retrieval in large-scale recommendations
75 citesProceedings of the 30th ACM international conference on information …, 2021
AAAI 2019
★ HIGHLIGHT
ATP: Directed Graph Embedding with Asymmetric Transitivity Preservation
69 citesThe Thirty-Third AAAI Conference on Artificial Intelligence (AAAI 2019), 2018
TL;DR: Preserves high-order directed reachability and transitivity via asymmetric non-linear dual-vector projections.
ACM TIST 2014
★ HIGHLIGHT
Vsrank: A novel framework for ranking-based collaborative filtering
66 citesACM Transactions on Intelligent Systems and Technology (TIST) 5 (3), 1-24, 2014
Proceedings of the 2017 ACM on web science conference, 151-160, 2017
TL;DR: Approximation algorithm resolving minimal feedback arc sets in web and noisy social hierarchies.
arXiv preprint arXiv:2203.02073, 2022
2022
★ HIGHLIGHT
Label leakage and protection from forward embedding in vertical federated learning
62 citesarXiv preprint arXiv:2203.01451, 2022
arXiv preprint arXiv:2106.05508, 2021
ICML 2023
★ HIGHLIGHT
Weak proxies are sufficient and preferable for fairness with missing sensitive attributes
44 citesInternational Conference on Machine Learning, 43258-43288, 2023
TL;DR: Proves that weak proxy representations achieve Pareto-optimal algorithmic fairness when sensitive attributes are unobserved.
arXiv 2024
★ HIGHLIGHT
Leveraging print debugging to improve code generation in large language models
40 citesarXiv preprint arXiv:2401.05319, 2024
TL;DR: Proposes LDB (Large Language Model Debugger), enabling LLMs to segment, trace, and verify variable execution states like human developers.
FL-ICML 2021
★ HIGHLIGHT
Defending against Reconstruction Attack in Vertical Federated Learning
40 citesFL-ICML 2021, 2021
Proceedings of the 21st ACM international conference on Information and …, 2012
2018
QDEE: Question Difficulty and Expertise Estimation in Community Question Answering Sites
35 citesThe International AAAI Conference on Web and Social Media (ICWSM 2018), 2018
Proceedings of the 21st ACM international conference on Information and …, 2012
2020
EndCold: An End-to-End Framework for Cold Question Routing in Community Question Answering Services
28 citesThe Twenty-Ninth International Joint Conferences on Artificial Intelligence …, 2020
NeurIPS W 2022
★ HIGHLIGHT
Netflix and forget: Efficient and exact machine unlearning from bi-linear recommendations
26 citesarXiv preprint arXiv:2302.06676, 2023
TL;DR: Provable exact machine unlearning algorithm removing user interaction data from bilinear recommendation models with minimal retraining.
ECML PKDD 2018
ColdRoute: effective routing of cold questions in stack exchange sites: J. Sun et al.
26 citesData mining and knowledge discovery 32 (5), 1339-1367, 2018
arXiv preprint arXiv:2301.07284, 2023
The Thirty-Seventh AAAI Conference on Artificial Intelligence (AAAI 2023 …, 2022
TL;DR: Formulates differentially private pairwise ranking loss optimization with DP-SGD for cross-silo advertising conversion.
Proceedings of the 32nd ACM International Conference on Information and …, 2023
TL;DR: Proposes perturbation mechanisms for intermediate representations to neutralize gradient inversion attacks in split networks.
arXiv preprint arXiv:2205.12412, 2022
Proceedings of the 18th ACM Conference on Recommender Systems, 819-822, 2024
TL;DR: Architects an end-to-end framework enabling multi-party ad attribution and CTR modeling within secure clean rooms.
Proceedings of the 29th ACM SIGKDD conference on knowledge discovery and …, 2023
TL;DR: Graph topological subsampling enabling extreme throughput and training speedup on massive recommender datasets.
arXiv preprint arXiv:2206.07998, 2022
NeurIPS ML Safety Workshop, 2022
The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI 2019), 2018
Proceedings of the Nineteenth ACM Conference on Recommender Systems, 940-944, 2025
TL;DR: Decouples query and pin entity embeddings to scale real-time retrieval across billions of candidates in Pinterest Ads.
The Ohio State University, 2019
US Patent 12,670,435, 2026
US Patent 12,651,208, 2026
US Patent 12,455,972, 2025
US Patent App. 18/865,581, 2025
US Patent 12,423,463, 2025
Patent
Data protection method, training method and apparatus for network structure, medium, and device
1 citesUS Patent App. 18/565,015, 2024
US Patent App. 18/539,851, 2024
US Patent App. 18/252,559, 2024
US Patent 11,763,204, 2023
US Patent 12,688,322, 2026
US Patent App. 18/865,611, 2025
US Patent App. 18/839,702, 2025
US Patent App. 18/623,714, 2024
US Patent App. 18/565,000, 2024
US Patent App. 17/953,255, 2024
US Patent App. 18/149,462, 2023
Patent
Diversifying recommendations by improving embedding generation of a graph neural network model
US Patent App. 17/985,788, 2023
US Patent App. 17/816,250, 2022