Siddharth Prasad

I am a researcher working on mechanism design for modern AI systems. I am currently a Research Assistant Professor at Toyota Technological Institute at Chicago (TTIC).

I received a PhD in Computer Science from Carnegie Mellon University, where I was advised by Nina Balcan and Tuomas Sandholm. During the PhD, I spent a summer working on recommender systems at Google Research. Before the PhD, I received a B.S. in math and computer science from Caltech.

My PhD Thesis: Mechanism Design and Integer Programming in the Data Age

• Inference Auctions
  with Keegan Harris, Asher Trockman, Nika Haghtalab, and Michael Jordan
  arXiv preprint 2026
• Auctions with Price Predictions
  with Muthu Sundar, Alec Sun, and Dravyansh Sharma
  arXiv preprint 2026
• In-Context Credit Assignment via the Core
  with Keegan Harris and Asher Trockman
  EC Workshop on Game Theory and Mechanism Design with LLMs 2026 [Spotlight talk]
• Parameter Tuning with Generalization Guarantees for GPU-Accelerated Linear Programming
  with Dravyansh Sharma
  arXiv preprint 2026
• Bicriteria Multidimensional Mechanism Design with Side Information
  with Nina Balcan and Tuomas Sandholm
  Mathematics of Operations Research 2026 (Supersedes our NeurIPS 2023 paper)
• Weakest Bidder Types and New Core-Selecting Combinatorial Auctions
  with Nina Balcan and Tuomas Sandholm
  AAAI Conference on Artificial Intelligence (AAAI) 2026 [Oral presentation]
• Revenue-Optimal Efficient Mechanism Design with General Type Spaces
  with Nina Balcan and Tuomas Sandholm
  arXiv preprint 2025
• New Sequence-Independent Lifting Techniques for Cover Inequalities and When They Induce Facets
  with Ellen Vitercik, Nina Balcan, and Tuomas Sandholm
  International Joint Conference on Artificial Intelligence (IJCAI) 2025
  [Best poster award (honorable mention) at MIP workshop 2024]
• Increasing Revenue in Efficient Combinatorial Auctions by Learning to Generate Artificial Competition
  with Nina Balcan and Tuomas Sandholm
  AAAI Conference on Artificial Intelligence (AAAI) 2025
  
• Bicriteria Multidimensional Mechanism Design with Side Information
  with Nina Balcan and Tuomas Sandholm
  Conference on Neural Information Processing Systems (NeurIPS) 2023
  [sigecom reading list]
• Content Prompting: Modeling Content Provider Dynamics to Improve User Welfare in Recommender Ecosystems
  with Martin Mladenov and Craig Boutilier
  RecSys Workshop on Causality, Counterfactuals, and Sequential Decision Making (CONSEQUENCES) 2023
  
• Structural Analysis of Branch-and-Cut and the Learnability of Gomory Mixed Integer Cuts
  with Nina Balcan, Tuomas Sandholm, and Ellen Vitercik
  Conference on Neural Information Processing Systems (NeurIPS) 2022 [Oral presentation]
  
• Maximizing Revenue under Market Shrinkage and Market Uncertainty
  with Nina Balcan and Tuomas Sandholm
  Conference on Neural Information Processing Systems (NeurIPS) 2022
  
• Improved Sample Complexity Bounds for Branch-and-Cut
  with Nina Balcan, Tuomas Sandholm, and Ellen Vitercik
  International Conference on Principles and Practice of Constraint Programming (CP) 2022
• Sample Complexity of Tree Search Configuration: Cutting Planes and Beyond
  with Nina Balcan, Tuomas Sandholm, and Ellen Vitercik
  Conference on Neural Information Processing Systems (NeurIPS) 2021 [Spotlight]
  
• Learning Within an Instance for Designing High-Revenue Combinatorial Auctions
  with Nina Balcan and Tuomas Sandholm
  International Joint Conference on Artificial Intelligence (IJCAI) 2021
  
• Efficient Algorithms for Learning Revenue-Maximizing Two-Part Tariffs
  with Nina Balcan and Tuomas Sandholm
  International Joint Conference on Artificial Intelligence (IJCAI) 2020
  
• Incentive Compatible Active Learning
  with Federico Echenique
  Innovations in Theoretical Computer Science Conference (ITCS) 2020
  
• Learning Time Dependent Choice
  with Zachary Chase
  Innovations in Theoretical Computer Science Conference (ITCS) 2019
• Walks on Primes in Imaginary Quadratic Fields