srujan@portfolio: ~ — tty1
Terminal session: srujan at portfolio, home directory, tty1last login: ——— on tty1
Computer Science + Economics + Math @ NC State
building systems, studying institutions, chasing useful work
available commands:
Hi, I'm Srujan Ponnapalli. I'm interested in systems — technical, financial, institutional — and the weird places where they fail.
Recently, that has meant research work with machine learning and chemistry data, student organization finance, data infrastructure, and projects around maps, misinformation, and public-interest technology.
I like work that makes messy information easier to understand or act on, whether that is a model, a dashboard, a map, or a clearer explanation to a question that stumps people (including myself).
Right now, I'm trying to become technically sharp enough to build serious things and intellectually broad enough to know what is worth building.
University of North Carolina · Chapel Hill, NC · Jun 2026 – Aug 2026
Shipped a Python desktop tool that automatically ingests new and changed files into iRODS, using PySide6 and Watchdog. Built its test suite from scratch and presented the design and release plan to RENCI leadership and iRODS Consortium members.
NC State University (AIMS Laboratory) · Raleigh, NC · Jan 2026 – Present
Developing physics-informed graph neural networks to predict PFAS persistence. Built an RDKit feature pipeline covering 14,000 chemical structures; implementing active learning and uncertainty quantification for environmental decision support.
University of North Carolina at Charlotte · Charlotte, NC · May 2024 – Sep 2024
Deployed a YOLOv7 object-detection model on a mobile robot, feeding on-device detections into its decision loop and demonstrating the system to technical stakeholders.
TypeScript · Next.js · scikit-learn · Agentic AI
1st Place @ WolfHacks 2026
Turning an explainable course check-in into a practical study plan, with the model's limits visible and the student in control.
Python · PySide6 · Open source
Turning research-data ingest into a hands-free desktop workflow, then building the theme, accessibility, and test foundations that help it hold up after launch.
TypeScript · Next.js · Amazon Bedrock
Grounding AI-assisted media verification in retrieved evidence by combining image analysis, fact-check databases, web search, and multiple model perspectives in one review flow.
[01] what are you studying?
CS, Econ, and Math. Call it either interdisciplinary ambition or a controlled intellectual spiral.
[02] what do you want to build?
Infrastructure, AI, governance, research, finance. In other words, systems that make other systems less stupid.
[03] are you more of a cs person or an econ person?
Computer science gives me the tools to build; economics reminds me that those tools eventually have to meet people, incentives, and reality.
[04] what kind of work are you looking for?
Technical work with institutional consequences: AI, data infrastructure, research tooling, finance, civic technology, or systems that help serious organizations operate better.
How machine learning systems become part of ordinary institutional life.
Using technical tools to understand economics, governance, behavior, and inequality.
Especially where models help researchers reason about complex systems.
Tools that help organizations with limited resources operate like they have real infrastructure.
How systems fail, how people trust them, and how to design against decay.