Systems Research Assistant
Amherst, MA
Skills
C / C++
Java
Python
LLM APIs
Docker and containers
Web Development
Interests
Studio Art
Ballroom Dance
Creative Writing
Mixology
Documents
Resume (View and download here)
Welcome to my webpage!
I first got my B.S. in
Chemical Engineering at Tufts University's School of Engineering and it was there that I
took introductory
CS courses and completely fell in love with computer science. I loved my Intro to
Computation course so much that
I decided to pursue computer science as a
second major.
During this time, I did computational biology research under Lenore Cowen as a Laidlaw
Scholar
and co-authored my first paper. Later, I joined my advisor, Jeffrey
Foster, in researching path-sensitive programming languages.
In 2023, I graduated from Tufts and enrolled in my current M.S./Ph.D. Computer Science
program at the University of Massachusetts Amherst. Here, I conducted LLM-based systems
research under Emery Berger that
ultimately culminated in my first projects, ChatDBG and Pythoness. I continue this work into
agentic software systems
as part of the LASER lab with my advisor, Yuriy Brun. My research focuses on the application
of LLMs to software engineering challenges
and specifically the integration of AI with traditional approaches and formal methods to
create hybrid
systems.
When I have it, I enjoy spending my free time ballroom dancing and cooking and mixing up
drinks for my friends.
I currently work in UMass Amherst's LASER Lab with my advisor, Yuriy Brun, investigating the use of LLMs and agentic tools and how their applications can make formal methods more accessible to software developers. Our work involves building hybrid systems that leverage both AI and formal methods to transform Python code and natural language specifications into formally verified software for greater code correctness assurances.
At UMass Amherst's PLASMA Lab, I worked with Emery Berger to investigate the use of LLMs as debugging tools. ChatDBG and CWhy are two such projects, and my research focused on modifying the former to take snapshots of the heap during program crashes and having ChatGPT converse with the debugger in order to suggest possible solutions to the user. ChatDBG is usable through a GDB, LLDB, and PDB interface with limited WinDBG extension capabilities. I presented the project and paper at FSE 2025 in Trondheim, Norway: https://dl.acm.org/doi/10.1145/3729355
Check out my ChatDBG presentation on YouTube!
Under my advisor, Jeffrey Foster, and alongside Ph.D. student Mark Aldrich, I worked on building RESTπ, a type inference engine that could automatically generate REST API specifications using path-sensitive type inference. Implemented for Ruby, RESTπ could generate API specifications with more exact and complete types by considering each distinct path through an endpoint implementation. I was involved in the early work for RESTπ, helping to develop the formalism behind the path-sensitive type inferencing, and later on, running some of the evaluation benchmarks for it as well. The paper was presented at SPLASH 2025 in Singapore: https://dl.acm.org/doi/10.1145/3763055
In the Cowen Lab under Lenore Cowen, I worked as an undergraduate research assistant as part of the Laidlaw Scholarship program, where I assisted Ph.D. student Mert Erden on developing ADAGIO, a graph-searching algorithm for protein-protein-interaction (PPI) networks to identify possible unidentified disease genes. This used known disease modules for neurological diseases such as Alzheimer's and Parkinson's to identify clusters of genes possibly linked to symptoms of these diseases. Our findings were presented at ACM-BCB 2022 in Chicago: https://dl.acm.org/doi/abs/10.1145/3535508.3545542
Completed a research internship with AWS as part of the Automated Reasoning team under Mike Hicks.
I was a teaching assistant for seven semesters, at both UMass Amherst and Tufts. At Tufts, I taught Discrete Mathematics under Prof. Karen Edwards and Cryptography under Prof. David Wittenberg, and at UMass Amherst, I was a TA for the Introduction to Computation course with Prof. David Barrington. This involved grading homework and exams, hosting office hours, and leading discussions, reviews, and workshops for hundreds of students.
At Harvard University, in the Widener Library Judaica Division, I helped develop new programming scripts and interfaces in FileMaker to make the library's digital collection both easier to navigate and faster to index. Many of my scripts made it easier to digitize records and compile the collection into visualized statistics and graphs.
I tutored remotely through Varsity Tutors in a wide array of subjects, such as programming in C, C++, and Python, web development, theory and mathematics, chemistry, engineering, and standardized test review. These were one-on-one private online sessions with students from all over the country, anywhere from 8th grade to adult learners, and usually involved reviewing course materials or creating and testing on my own custom materials.
I interned at Tortoise Media with their Intelligence Team as part of the Laidlaw Scholars program. This involved collecting research and information for the company's stories and website pages, specifically a project to compile MP financial transactions into an interactive graph that could be queried. I analyzed this data and wrote a clustering algorithm to help the program distinguish MPs and donors with similar names.
M.S. / Ph.D. Computer Science
B.S. Chemical Engineering and Computer Science