I'm 0 years old.

I studied at CMU's School of Computer Science.

GitHubLinkedInEmail
Scroll to see more

I like to work in this intersection!

But around it is interesting, too.

MachineLearningSystemsEngineeringProductDevelopment

Experience

Fundusol logo

Co-founder, ML/Software Lead

Fundusol

October 2021 - Present

Remote

Co-founded Fundusol and worked remotely for several years on a suite of Agrivoltaic system simulators during school. Led a team to develop 3D visualizations, GPU-based ray tracing software, and genetic algorithms for designing and optimizing agricultural PV systems. Pitched to VCs, pitch competitions, and engaged with solar developers, which helped secure $100k in revenue and $100k+ in funding (from Stanford TomKat Center and the U.S. Department of Energy). Work was done mostly in Python (PVLib and Streamlit) and C++. Learned about ML for Agriculture, CUDA for ray tracing / irradiance modeling, and product design and development.

Citadel Securities logo

Software Engineering Intern

Citadel Securities

June 2024 - August 2024

Chicago, IL

Tested, benchmarked, and optimized a low-latency trading application. Achieved a 75% reduction in latency which unlocked thousands of dollars in daily trading opportunity. Learned about distributed systems, networking, and software engineering practices in a high-risk, high-speed environment.

Amazon Web Services logo

Software Engineering Intern

Amazon Web Services

June 2023 - September 2023

Seattle, WA

Developed a patching and vulnerability detection tool that scanned 10,000 EC2 servers per day. Created a tight iteration loop with the ELB team and improved their workflow, eventually saving them over 10 hours per week with the tool. Learned about product design, web development, and the AWS SDK. Used Python and pure JS/HTML/CSS (so sad).

Gallus Insights logo

Software Engineering Intern

Gallus Insights

June 2022 - August 2022

Remote

Led development of Gallus Analytics, a full-stack mortgage data visualization tool, and created a predictive model for mortgage market metrics. The tool secured hundreds of customers and helped win a multibillion-dollar lender as a client. Learned about product design, full-stack development, and data visualization.

Codellege by Softtek logo

Coding Instructor

Codellege by Softtek

Summer 2020 & Spring 2021

Monterrey, Mexico

Designed 3-month Arduino programming course and taught to dozens of underserved students as part of the "START" program.

Projects

Avalon Learn screenshot

Avalon Learn

January 2025 - Present

Reinventing grading and student feedback in higher education with a new platform. In just 3 months, we built a tool that is being used by several CMU / Pitt courses in Spring 2025, and secured 20+ more commitments for the Summer and Fall of 2025. We've been awarded $12,500 through two pitch competitions at CMU (CMU VC and McGinnis).

Grand Prize at CMU VC ($5,000)
Winner of McGinnis Social Enterprise Track ($7,500)
COMPASS screenshot

COMPASS

October 2024 - December 2024

We implemented an OMPT performance analysis and debugging tool for C++ programs with OpenMP directives. COMPASS performs event logging and deadlock detection at runtime, and outputs several visualizations post-mortem: an execution graph as well as a detailed breakdown of time spent across critical regions, parallel work, barrier waiting, and other key areas.

1st Place Project in CMU's 15-418: Parallel Computer Architecture and Programming (Fall 2024)
Needle Weld screenshot

Needle Weld

October 2024 - December 2024

Developed an operator tree GPU kernel fusion library for the Needle machine learning library. Needle Weld scans computational graphs at runtime, detects "trees" of binary operators in the graph (that would have previously each been independently executed by separate kernels in topological order). It then generates and compiles a CUDA kernel to combine those operations and replaces that tree with the new kernel in the computational graph, all at runtime.

Automating Codenames Spymasters screenshot

Automating Codenames Spymasters

March 2024 - May 2024

Developed a novel ensemble word-to-vector model to create an automated "spymaster" for the popular game of "Codenames". Resulted in a 27% improvement from the state-of-the-art. Novel techniques included a multi-definition / mixture of models approach and use of the therandomized weighted majority algorithm.

TAI screenshot

TAI

February 2023 - February 2024

Developed an LLM question-answering app for YouTube videos that got 1st place at CMU's TartanHacks 2023. Expanded to a Chrome Extension that got over 500 downloads and a feature on Chrome Web Store. Later, partnered with a leading Sri Lankan educational platform to expand the tool's reach.

1st Place at TartanHacks 2023
Featured on Chrome Web Store
500+ downloads
Panda Bot screenshot

Panda Bot

September 2023 - November 2024

Collaborated with Professor Ada to develop a chain-of-thought, multi-agent RAG system to help instructors answer student questions more efficiently in the notoriously difficult CMU Theoretical Computer Science course, 15-251. Answered thousands of questions more quickly and effectively with Panda Bot.

Point & Prove screenshot

Point & Prove

2023-2024

Created a web app for teaching constructive logic and was invited by Professor Marijn Heule to present to CMU's course on Logic and Mechanized Reasoning, 15-311