Kartik Sharma

searching for my blood flow

prof_pic2.jpg

🦺 🤖 🎛️

I am a final-year Ph.D. candidate at Georgia Institute of Technology, being advised by Prof. Srijan Kumar and mentored by Dr. Rakshit Trivedi. Through my research, my goal is to study and build machine learning (🤖) models, which are robust to perturbations (🦺) and can be easily controlled (🎛️) by the users, specifically, in relational (eg., graph), ordered (eg., language, decision process), and dynamic (eg., dynamic graph, multi-agent) environments. These environments are unique in how they can model numerous real-world applications while providing a mathematical formalism. My motivation arises from the fact that these models are inherently probabilistic and non-deterministic by nature and may behave differently from the users’ expectations, hindering their widespread adoption. Thus, it is important that we ensure that the model’s behavior remains stable under reasonable perturbations, but changes precisely as or when directed through explicit controls.

I have had the pleasure of getting valuable industrial experience throughout my Ph.D.: working with Amazon RISC (Regulatory Intelligence, Safety and Compliance) Science team in 2025, Microsoft Research for industry group in 2024, and Visa Research in 2024. Before this, I graduated from IIT Delhi in 2021, where I was fortunate to be advised by Prof. Sayan Ranu and mentored by Prof. Sourav Medya.

news

Sep 19, 2025 Paper accepted at NeurIPS 2025 (spotlight)! Paper and code coming soon!
Sep 19, 2025 Two papers accepted at EMNLP 2025 (main and findings)!
Aug 16, 2025 Finished my internship with the Regulatory Compliance Science team at Amazon.
May 14, 2025 Successfully proposed my thesis and now officially a PhD candidate!
May 14, 2024 Spending my summer at MSR, Redmond in the Research for Industry Lab.

research overview

Preview of research overview

đź”— Open full PDF

selected publications

Enabling user-controllable generation.

  1. Diffuse, Sample, Project: Plug-And-Play Controllable Graph Generation
    Kartik Sharma, Srijan Kumar , and Rakshit Trivedi
    In Forty-first International Conference on Machine Learning , 2024
  2. OG-RAG: Ontology-grounded retrieval-augmented generation for large language models
    Kartik Sharma, Peeyush Kumar , and Yunqing Li
    In The 2025 Conference on Empirical Methods in Natural Language Processing , 2025
  3. Inner Speech as Behavior Guides: Steerable Imitation of Diverse Behaviors for Human-AI coordination
    Rakshit Trivedi* , Kartik Sharma*, and David C. Parkes
    In The Thirty-ninth Annual Conference on Neural Information Processing Systems , 2025

Evaluating and enhancing the robustness.

  1. Task and model agnostic adversarial attack on graph neural networks
    Kartik Sharma, Samidha Verma , Sourav Medya , and 2 more authors
    In Proceedings of the AAAI Conference on Artificial Intelligence , 2023
  2. Temporal dynamics-aware adversarial attacks on discrete-time dynamic graph models
    Kartik Sharma, Rakshit Trivedi , Rohit Sridhar , and 1 more author
    In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 2023
  3. Sysformer: Safeguarding Frozen Large Language Models with Adaptive System Prompts
    Kartik Sharma, Yiqiao Jin , Vineeth Rakesh , and 4 more authors
    arXiv preprint arXiv:2506.15751, 2025
  4. Personalized Layer Selection for Graph Neural Networks
    Kartik Sharma, Vineeth Rakesh , Yingtong Dou , and 2 more authors
    Transactions on Machine Learning Research, 2025