cv

Researcher

Momentum Lab @ CMU RI (led by Jeff Ichnowski)

Understanding textile deformation using ML for use in dynamic manipulation such as flinging and flattening.

• Computer Vision and Robotics
• Manipulation

Researcher

Biorobotics Lab @ CMU RI (led by Howie Choset)

Building ML system for Pipe Mapping robot to detect and segment metal anomalies. Currently using generative methods to create a more diverse training set for segmentation methods.

• Computer Vision and Robotics
• GAN + Diffusion
• Image Segmentation and Generation

Researcher

Search Based Pathfinding Lab @ CMU RI (led by Max Likhachev)

Researching usability of Machine Learning to generate better and faster results for multi-agent pathfinding questions (e.g. applicable for finding paths for robots in warehouses or self-driving cars).

• A*
• Pathfinding
• MAPF
• EECBS
• LaCAM

Research Scholar

New York University's Center for Cybersecurity (w/ Nasir Memon)

Developing a CAPTCHA-like technology for identifying video and voice deepfakes (paper in preparation) using machine learning models using (among other packages) nnabla, librosa on an HPC.

Research Intern

Computer Science and Engineering, NYU

Identified manipulated images and false statements made by politicians with Reverse Image Search and drafted candidate algorithm to improve Reverse Image Search, specifically for robustness against manipulations.

Intern

Amber Semi, Inc

Leveraged existing router network infrastructure, created method and proof-of-concept to associate user MAC addresses with user contact information and web-browsing cookies to improve personalization and co-developed patent for Privacy and the Management of Permissions (patent under application).

Undergraduate

Carnegie Mellon University

Statistics & Machine Learning and Computer Science

• Computer Vision (PhD level)
• Visual Learning and Recognition (PhD level)
• Deep Learning (Masters level)
• Parallel & Sequential Algorithms
• Computer Systems
• Functional Programming
• Imperative Computation
• Probability and Statistical Inference
• Statistical Graphics and Visualization

High School

The Harker School

Dean's List High Honors

Dean's List

1st Place Coolest Graphs

CMU Statistics Department

Manrattan - A Look into NYC's Rats (link)

Work Smarter Not Harder: Simple Imitation Learning with CS-PIBT Outperforms Large Scale Imitation Learning for MAPF

International Conference on Robotics and Automation (Submitted to ICRA 2025)

Multi-Agent Path Finding (MAPF) is the problem of effectively finding efficient collision-free paths for a group of agents in a shared workspace. The MAPF community has largely focused on developing high-performance heuristic search methods. Recently, several works have applied various machine learning (ML) techniques to solve MAPF, usually involving sophisticated architectures, reinforcement learning techniques, and set-ups, but none using large amounts of high-quality supervised data. Our initial objective in this work was to show how simple large scale imitation learning of high-quality heuristic search methods can lead to state-of-the-art ML MAPF performance. However, we find that, at least with our model architecture, simple large scale (700k examples with hundreds of agents per example) imitation learning does extit{not} produce impressive results. Instead, we find that by using prior work that post-processes MAPF model predictions to resolve 1-step collisions (CS-PIBT), we can train a simple ML MAPF model in minutes that dramatically outperforms existing ML MAPF policies. This has serious implications for all future ML MAPF policies (with local communication) which currently struggle to scale. In particular, this finding implies that future learnt policies should (1) always use smart 1-step collision shields (e.g. CS-PIBT), (2) always include the collision shield with greedy actions as a baseline (e.g. PIBT) and (3) motivates future models to focus on longer horizon / more complex planning as 1-step collisions can be efficiently resolved.

AI-assisted Tagging of Deepfake Audio Calls using Challenge-Response

arXiv (submitted to AsiaCCS 2025)

Scammers are aggressively leveraging AI voice-cloning technology for social engineering attacks, a situation significantly worsened by the advent of audio Real-time Deepfakes (RTDFs). RTDFs can clone a target's voice in real-time over phone calls, making these interactions highly interactive and thus far more convincing. Existing literature has largely been ineffective against RTDF threats. We introduce a robust challenge-response-based method to detect deepfake audio calls with high accuracy.

Improving Learnt Local MAPF Policies with Heuristic Search

International Conference on Automated Planning and Scheduling (ICAPS 2024)

We improve a ML local policy by using heuristic search methods on the output probability distribution to resolve deadlocks and enable full horizon planning for Multi Agent Planning (MAPF). We show several model-agnostic ways to use heuristic search with ML that significantly improves the local ML policy's success rate and scalability. To our best knowledge, we demonstrate the first time ML-based MAPF approaches have scaled to similar high congestion as state-of-the-art heuristic search methods.