Hi! I'm Shreeya Khurana.

• Interned on the ML Exploration team under the ML Infrastructure Organization
• Built an interactive debugger on ML workflow orchestration platform, saving MLEs 20 min per iteration cycle
• Performed prompt engineering and benchmarked LLMs for query intent classification, and gauged performance-cost trade-offs

• Developing a model which learns a driver's situational awareness through eye-gaze tracking.
• Constructed a driving dataset with continuous per-object situational awareness labels, traffic agent states, and driver eye-gaze collected using a protocol in a VR driving simulator
• Developed a gaze-based ML model of a driver’s per-object situational awareness in various traffic scenarios
• Paper accepted for presentation at 2024 Conference on Robot Learning (CoRL)

• Built a Python package for input validation and manipulation for simulation of a robotic fulfillment center
• Reduced simulation time significantly by catching errors in 48GBs of required data before a simulation job is launched
• Improved interpretability and user experience of simulation tooling and reporting

• Teaching Assistant on the course staff for CMUs Introduction to Machine Learning class for students in the School of Computer Science (10-315).
• Developed and taught weekly recitations covering lecture material for that week and walking students through relevant practice problems.
• Hosted weekly office hours to help students with homework and conceptual questions, graded homework, and developed and graded exam questions

• Implemented NPC (non-player character) intent conditioning in trajectory prediction ML model
• Trained, deployed, and tested model on internal simulation dataset, achieving >5% increase in collision recall

• Teaching Assistant on the course staff for CMUs Concepts in AI class (07-180).
• Taught Recitations, held weekly office hours, wrote and graded homework and exam problems.

• Improved upon a system for AI mediated collaboration in hands-on craft learning with computer vision

• Improved upon system for Video Object Tracking with Deep Reinforcement Learning

• Developed an algorithm for site-selection in autonomous swarms for use in humanitarian assistance and disaster-relief missions
• First author paper accepted for presentation and publication in the Proceedings of the 12th International Conference on Agents and Artificial Intelligence (ICAART 2020) - all papers accepted to ICAART 2020 were peer reviewed by at least two experts from the International program committee, in a double-blind review process.

• Developed an algorithm to analyze large datasets using parallel computing
• Identified problems in collaboration tools used by scientists at JHU-APL and recommended solutions

• Compared different models available for prediction of magnetopause location changes
• Invited speaker at the 9th Community Coordinated Modeling Center Workshop held at College Park, Maryland (April 2018)

• Worked on Geomagnetic Storms: A Study of Relationship between Geomagnetic Storms and the Interplanetary Magnetic Field, and Monitoring Geomagnetic Storms in the Ionosphere with GPS Errors
• Invited speaker at the the 8th Community Coordinated Modeling Center Workshop held in Annapolis, MD in 2016
• Poster presentation: Geospace Environment Modeling Summer Workshop, Portsmouth, Virginia (June 2017)

Modeling Drivers' Situational Awareness from Eye-Gaze for Intelligent Alerts (2024) – paper accepted for CoRL 2024

S. Khurana and D. Sofge. (2020). Quorum Sensing Re-Evaluation Algorithm for N- Site Selection in Autonomous Swarms. In Proceedings of the 12th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART, 193-198. https://www.scitepress.org/Papers/2020/89578/89578.pdf

Collado-Vega, Y. M., Dredger, P., Lopez, R. E., Khurana, S., Rastaetter, L., Sibeck, D., & Anastopulos, M. (2023). Magnetopause standoff position changes and geosynchronous orbit crossings: Models and observations. Space Weather, 21, e2022SW003212. https://doi.org/10.1029/2022SW003212

Dredger, P. M., Lopez, R. E., Collado-Vega, Y. M., Khurana, S., & Rastaetter, L. M. (2023). Investigating potential causes for the prediction of spurious magnetopause crossings at geosynchronous orbit in MHD simulations. Space Weather, 21, e2022SW003266. https://doi.org/10.1029/2022SW003266

My areas of expertise are computer science, specifically AI and ML. Here is a summary of the languages and libraries I have worked with.

• Python
• NumPy
• Tensorflow
• PyTorch
• OpenCV
• C
• Java
• C++
• MATLAB
• R
• HTML, CSS, JavaScript