Sai Koti Reddy Danda

Research

My research spans multi-agent systems, agentic AI, large language models (LLMs), and reinforcement learning, including constrained and risk-averse multi-agent RL, with foundations in game theory and stochastic optimization. I have carried it into products: RL-driven autoscaling and GPU orchestration at Avesha, and multi-agent process intelligence for banking at kAIgentic.

Google Scholar
Download CV

Funding

Research grants

IBM Research Math Council

Making Game Theory Practical for Decision-Making in Multi-Agent Systems

Five-year IBM Research Math Council grant I co-wrote at IBM Research. It funded research in constrained and risk-averse multi-agent RL, with publications at AAMAS 2019 and 2021, IEEE SCC 2022, and ECAI 2023.

DST · Robert Bosch Centre

Distributed Multi-Agent Algorithms for Dynamic Control of Microgrids

Five-year grant I co-wrote, funded by the Department of Science & Technology, India and the Robert Bosch Centre for Cyber-Physical Systems, IISc. It supported publications at IEEE SmartGridComm, IEEE ISGT-Europe, IEEE CDC, and AAMAS 2021.

Papers

Publications

  1. Truthful and Equitable Lateral Transshipment in Multi-Retailer Systems
    G. Shakya, S. Danda, S. Nath, P. Dayama, and S. S. K. Sajja
    ECAI 2023, 26th European Conference on Artificial Intelligence · link
  2. Inventory Pooling using Deep Reinforcement Learning
    K. Sampath, S. Nishad, S. Danda, P. Dayama, and S. Sankagiri
    IEEE International Conference on Services Computing (SCC), 2022
  3. Attention Actor-Critic Algorithm for Multi-Agent Constrained Co-operative Reinforcement Learning
    P. Parnika, R. B. Diddigi, S. Danda, and S. Bhatnagar
    AAMAS 2021 · joint first author · link
  4. Atomic Execution of Optimization Transactions on Permissioned Blockchains
    K. Narayanam, K. Sampath, K. Kumar, S. Danda, and P. Dayama
    IEEE International Conference on Blockchain and Cryptocurrency (ICBC), 2021
  5. Spot Collaborative Shipping Sans Orchestrator using Blockchain
    K. Sampath, S. Danda, K. Kumar, K. Narayanam, P. Dayama, and S. Sankagiri
    IEEE International Conference on Blockchain, 2020
  6. Invoice Reconciliation using Blockchain
    K. Sampath, P. Dayama, S. Danda, and V. D. Pandit
    INFORMS Annual Meeting, 2020 (presentation) · link
  7. Collaborative Shipping Sans Orchestrator using Blockchain
    K. Sampath, S. Danda, P. Dayama, and V. D. Pandit
    INFORMS Annual Meeting, 2020 (presentation) · link
  8. Stochastic Game Frameworks for Efficient Energy Management in Microgrid Networks
    S. Nayak, C. A. Ekbote, A. P. S. Chauhan, R. B. Diddigi, P. Ray, A. Sikdar, S. Danda, and S. Bhatnagar
    IEEE PES ISGT-Europe, 2020 · link
  9. Risk Averse Reinforcement Learning for Mixed Multi-Agent Environments
    S. Danda, A. Saha, S. G. Tamilselvam, P. Agrawal, and P. Dayama
    AAMAS 2019
  10. Actor-Critic Algorithms for Constrained Multi-Agent Reinforcement Learning
    R. B. Diddigi, S. Danda, K. J. Prabuchandran, and S. Bhatnagar
    AAMAS 2019 · joint first author · link
  11. A Unified Decision Making Framework for Supply and Demand Management in Microgrid Networks
    R. B. Diddigi, S. Danda, K. Narayanam, and S. Bhatnagar
    IEEE SmartGridComm, 2018 · joint first author · link
  12. Generalized Deterministic Perturbations for Stochastic Gradient Search
    K. Chandramouli, K. J. Prabuchandran, S. Danda, and S. Bhatnagar
    IEEE Conference on Decision and Control (CDC), 2018
  13. Improved Hessian Estimation for Adaptive Random Directions Stochastic Approximation
    S. Danda, L. A. Prashanth, and S. Bhatnagar
    IEEE Conference on Decision and Control (CDC), 2016 · link
  14. An Improvement to k-Nearest Neighbor Classifier
    S. Danda, T. H. Sarma, P. Viswanath, and S. Raghava
    3rd International Conference on Data Management (ICDM), 2010

Intellectual property

Patents

  • Controlling Placement of Workloads of an Application within an Application Environment · US11805073B2 · granted
  • Data Quality Control · US11169976B2 · granted
  • Distributed Service Delivery System with Adaptive Autoscaling of Compute Resources · filed
  • Service Network Employing Reinforcement Learning-Based Global Load Balancing · filed
  • Multi-Objective Driven Refactoring of Monolith Application Using Reinforcement Learning · US20230177337A1
  • Decentralized Online Multi-Agent Visual Question Answering · US20220188362A1
  • Factchecking Artificial Intelligence Models Using Blockchain · US20220237477A1
  • Option-Based Distributed Reservation System · US20210158229A1
Defensive publications
  • System and Method for Executing Optimization Transactions in Distributed Ledger Technology Based Applications
  • System and Method for Processing State Dependent Optimization Transactions on Blockchain
  • Layer-wise Model Compression/Acceleration with Multiple Objectives
  • Hierarchical Multi-agent Systems for Multi-objective / Multi-metric Classification
  • Behavioral Constraints Inclusive Conversation Goal Evaluation
  • System and Method for Blockchain Enabled Market Place for Real-time Analytical Models
  • Outcome Based Transaction Reconciliation
  • System and Method for External Recommendation Network Based Review & Incentivization of Research Reviews Using Blockchain
  • System and Method for Geographic Proximity Based Contact Recommendation for Meeting