Jugal Gajjar
Graduate student at The George Washington University working on reasoning-centric AI systems with a focus on Large Language Models, trustworthy and explainable AI, and AI for code intelligence and security.
I also explore applications in code intelligence and software security, where I combine graph learning and LLMs to build systems for vulnerability detection, reasoning, and automated repair.
My broader goal is to develop trustworthy AI systems that can collaborate effectively with humans in high-stakes environments. Learn more →
RSAT
Structured attribution framework for training small LLMs to generate step-by-step reasoning with explicit citations, improving faithfulness and interpretability.
Adaptive RAG
Reinforcement learning-based retrieval framework that learns when and how much to retrieve, improving efficiency and reasoning in LLM systems.
VulnGraph
Graph + LLM framework combining structural and semantic reasoning for explainable vulnerability detection in source code.
- RSAT: Structured Attribution Makes Small Language Models Faithful Table Reasoners — Under Review at SURGeLLM Workshop @ ACL 2026 [Link]
- Adaptive RAG: Learning When and How Much to Retrieve Across Model Scales — Under Review GEM Workshop @ ACL 2026 [Link]
- Feature Attribution Stability Suite (FASS): How Stable Are Post-Hoc Attributions? — Under Review XAI4CV Workshop @ CVPR 2026 [Link]
-
Graduate Assistant — The George Washington University (Sep. 2025 — Present)
- Supported graduate-level Big Data and Analytics (CSCI 4907/6444) with focus on distributed systems and machine learning.
- Guided students on Spark, Hadoop, AWS S3, and programming in Python, Java, and Linux.
-
Teaching Assistant — Navrachana University (Jan. 2024 — May 2024)
- Mentored 80+ students on software testing, debugging, and technical writing through structured feedback.
- Contributed to curriculum delivery and assessment for test-driven development.
-
NVIDIA Jetson AI Project Coordinator — Navrachana University (Mar. 2022 — Jun.
2022)
- Led workshops on computer vision and NLP with Jetson Nano for real-time embedded AI applications.
- Guided student research projects recognized by the NVIDIA Deep Learning Institute.
Email: jugal.gajjar@gwu.edu