Biography

Hi, welcome to my homepage. I am currently a tech-lead, senior AI/ML Scientist in Amazon Ads AI Foundation team focusing on building general-purpose foundation models with strong predictive and reasoning abilities by learning from large-scale customer journey history of Advertising and retail events. I was a Principal AI scientist at Keystone AI where I lead a small team building cutting-edge AI-driven prediction and decision science for real-world business operations. I spent 5+ years in Forecasting Science team of Supply Chain Optimization Technology at Amazon, where I was the tech lead for team’s AI research and deployment of large-scale deep learning models for direct revenue-impacting supply chain demand forecasting. I started my career at Bloomberg as a quantitative analyst developing Cross-Asset Derivative Pricing products and also worked for Equity Option Trading desk at Morgan Stanley. I got my Ph.D. from Applied Math and Statistics Department of Johns Hopkins University.

I am a passionate researcher, builder and practitioner of real-world AI/ML models for problems such as Supply Chain Optimization and Quantitative Trading. My current research interests include

• Post-training of LLMs
• Deep Learning architecture innovations for foundation models
• Multimodal forecaster with LLM and time series

Publications

• TAT: Temporal-Aligned Transformer for Multi-Horizon Peak Demand Forecasting
  KDD 2025 Workshop on AI for Supply Chain
  Zhiyuan Zhao, Sitan Yang, Kin G. Olivares, Boris N. Oreshkin, Stan Vitebsky, Michael W. Mahoney, B. Aditya Prakash, Dmitry Efimov
• RSight: A deep neural network for product demand forecasting over geographic regions
  KDD 2025 Workshop on AI for Supply Chain
  Hanjing Zhu, Sitan Yang, Tanmay Gupta, Harrison Waldon, Kin G. Olivares, Rahul Gopalsamy
• SPADE-S: A Sparsity-Robust Foundational Forecaster
  KDD 2025 Workshop on AI for Supply Chain
  Malcolm Wolff, Matthew Li, Ravi Kiran Selvam, Hanjing Zhu, Kin G. Olivares, Ruijun Ma, Abhinav Katoch, Shankar Ramasubramanian, Mengfei Cao, Roberto Bandarra, Rahul Gopalsamy, Stefania La Vattiata, Sitan Yang, Michael W. Mahoney
• GNN-Enhanced Meta-Learning for Peak Period Demand Forecasting with Proxy Data
  WACV 2024 Workshop on Physical Retail AI
  Zexing Xu, Linjun Zhang, Sitan Yang, Rasoul Etesami, Hanghang Tong, Huan Zhang, Jiawei Han
• GEANN: Scalable Graph Augmentations for Multi-Horizon Time Series Forecasting
  KDD 2023 Workshop on Mining and Learning with Graphs
  Sitan Yang, Malcolm Wolff, Shankar Ramasubramanian, Vincent Quenneville-Belair, Ronak Metha, Michael W. Mahoney
• MQRetNN: Multi-Horizon Time Series Forecasting with Retrieval Augmentation
  KDD MiLeTS 2022 (Best ML Paper in Amazon Consumer Science Summit 2022)
  Sitan Yang, Carson Eisenach, Dhruv Madeka

Talks

• Multi-horizon Time Series Forecasting with Retrieval Augmentation
  Bloomberg Quant Seminar, Oct. 2022
  Bloomberg Quant Seminar is a premier seminar series that takes place in New York and covers a wide range of topics in quantitative finance and technology, chaired by Bruno Dupire, head of Quantitative Research at Bloomberg LP.