I am a Ph.D. candidate at the Georgia Institute of Technology, working at the Institute for Robotics and Intelligent Machines (IRIM). I am advised by Prof. Matthew Gombolay.
I focus on natural language processing and machine learning, with an emphasis on improving Large Language Models (LLMs) for generation and refinement. My long-term goal is to create intelligent systems that help engineers, developers, and designers work more efficiently.
Research Directions
- Learning from Feedback: Building AI systems that improve through verification, evaluation, and iterative feedback (ICLR’25, JSP’25).
- Learning to Reason Under Constraints: Enabling AI systems to reason explicitly under physical, safety, and task-specific constraints (Preprint).
- Learning Which Feedback to Trust: Developing AI systems that assess feedback quality and learn to utilize reliable feedback sources effectively.
- Learning from Synthetic Data: Building AI systems that learn effectively from synthetic or unrealistic data by identifying high-quality, representative examples that improve real-world performance (SANER’24).
Selected papers
- Generating CAD Code with Vision-Language Models for 3D Designs: ICLR 2025
- Learning Defect Prediction from Unrealistic Data: SANER 2024
- Can LLMs Patch Security Issues?: JSP 2025
- Constraints-of-Thought: A Framework for Constrained Reasoning in Language-Model-Guided Search: Under review
Contact me
- Google Scholar
- Email: kalrashedy3@gatech.edu
- Github