Bhavik Agarwal

I'm a Machine Learning Research Scientist at MasterControl and recent M.S. graduate in Computer Science from Johns Hopkins University (advised by Prof. Casey Overby Taylor). My research focuses on Small Language Models, Multi-Agent Systems, and structured generation for trustworthy AI in safety-critical workflows.

Prior to my master's, I was a research mentee at Google DeepMind and worked as an engineer at BlueStacks and SAP Labs. I received my B.Tech. in Computer Science from IIIT Delhi.

Email / CV / Google Scholar / LinkedIn

Research

I work on trustworthy AI for regulated industries (life sciences, healthcare), focusing on multi-agent systems, knowledge graphs, and structured generation with verifiable outputs. Current interests include GPU optimization for LLM training/inference at the kernel level.

News

[January, 2026] Received EACL 2026 Strong Accept for our BMR digitization paper!
[November, 2025] Presented our RAGulating Compliance paper at ISWC 2025 in Japan!
[June, 2025] Patent from SAP internship accepted! Also presented at PDA Good Digital Manufacturing Conference, Switzerland.
[May, 2025] Successfully defended my Master's Thesis at Johns Hopkins University!
[April, 2025] ThinkJSON highlighted as Hugging Face Spotlight Paper and covered by MasterControl's GxP Lifeline.
[December, 2024] Paper on synthetic patient messages published in Journal of Biomedical Informatics and featured by JHU BME.
[July, 2024] Selected as ACM Tapia Scholar for Richard Tapia Conference.
[August, 2023] Named recipient of Google DeepMind CS Research Mentorship Program.

Publications & Patents

See my Google Scholar profile for the complete and most recent publications.

📄	From Paper to Structured JSON: An Agentic AI Workflow for Compliant BMR Digital Transformation Bhavik Agarwal, Nidhi Bendre, Viktoria Rojkova EACL 2026 arXiv

🤖	Think Inside the JSON: Reinforcement Strategy for Strict LLM Schema Adherence Bhavik Agarwal, Ishan Joshi, Viktoria Rojkova Under review at ACL 2026 \| Hugging Face Spotlight Paper arXiv

🕸️	RAGulating Compliance: A Multi-Agent Knowledge Graph for Regulatory QA Bhavik Agarwal, Hemant Sunil Jomraj, Simone Kaplunov, Jack Krolick, Viktoria Rojkova ISWC 2025 paper

🎓	Design and Evaluation of GPT and RAG-Based QA Systems for Genomic and Pharmacogenomic Patient Queries Bhavik Agarwal Master's Thesis, Johns Hopkins University, 2025 thesis

⚖️	Hierarchical Agentic Retrieval and Reasoning System for Enhanced Query-Scenario Routing in SAP's Enterprise AI Platforms Bhavik Agarwal, Julien Yu, Anil Babu Ankisettipalli, Sebastien Schrieber, Rebecca Danford, Aarti Arikatala US Patent US12493636B1 patent

⚖️	Multi-Agent Knowledge Graph Retrieval with Evidence Grounding Bhavik Agarwal, Hemant Sunil Jomraj, Simone Kaplunov, Jack Krolick, Viktoria Rojkova US Patent Application 19/355,798

🏥	Taxonomy-based prompt engineering to generate synthetic drug-related patient portal messages Natalie Wang, Sukrit Treewaree, Ayah Zirikly, Yuzhi L. Lu, Michelle H. Nguyen, Bhavik Agarwal, Jash Shah, James Michael Stevenson, Casey Overby Taylor Journal of Biomedical Informatics, 2024 paper

📊	Time Series Nowcasting of India's GDP with Machine Learning Nimisha Malik, Bhavik Agarwal ICAIoT, 2022 paper

🔬	Exploration of Negative Mass as a Harbinger of Dark Energy Aryaman Babbar, Bhavik Agarwal IOSR Journal of Applied Physics, 2020 paper

Website design inspired by Jon Barron