Priyam Kabra
I’m Priyam, a CS, Math, & Robotics student at the University of Maryland. I like building things where software meets the physical world.
This summer I was a software engineering intern at Verisign, automating reporting for the .com and .net registries. During the year I work on Jupiterp, a course planner used by 40,000+ UMD students, and TA CMSC330 (OCaml, Rust, Lambda Calc).
Experience
- University of Maryland — Teaching Assistant, CMSC330
Lead a discussion section and hold office hours for Organization of Programming Languages (OCaml, Rust, lambda calculus).
- Verisign — Software Engineering Intern
Automated weekly reporting for the operator of the .com and .net registries. Upgraded internal PMO reporting, automated monitoring-compliance checks and follow-ups, and prototyped a big-data migration for SLA reporting.
- Jupiterp — Software Engineer, Core Contributor
Open-source course planner for UMD students. Built the custom calendar events feature. Currently helping migrate PlanetTerp professor and grade data into Jupiterp.
- Rip Current Sports — Software Engineer, contract
Built an internal tool for a swim-equipment company through Project LIFT: automated sales-lead discovery, plus a RAG assistant that drafts grant applications from past ones. FastAPI, Supabase, Gemini, Claude.
- OneClique (UMD ACM) — Software Engineering Lead
Led a 20-person team building an AI assistant for UMD classes, dining, and transit. Headed the scheduling subteam, which generates and ranks class schedules from plain-English requests.
Education
- University of Maryland — B.S. Computer Science & Mathematics, minor in Robotics
4.0 GPA. Startup Shell, AppDev Club, Project LIFT. National Merit Scholarship winner and Eagle Scout.
Projects
Distributed compute in the browser. Phones and laptops join a cluster by scanning a QR code; jobs are split across Web Workers and re-queued if a device drops. Adding a second device halved a 35-second job. Built at YHack 2026.
Wearable motion capture. Two arm-mounted ESP32 IMUs and an iPhone’s LiDAR capture a 3D scene and skeleton in real time and replay it; limbs without sensors are filled in with inverse kinematics. 2nd place, app development track, Bitcamp 2026.
A Braille printer for under $200, compared to $10–15k for commercial embossers. A Raspberry Pi drives a three-axis punch and paper feed at over 99% embossing accuracy, with software that converts text, images, and audio to Braille. Presented at the NJ Center for Deafblindness conference.
A 3D renderer written from scratch: projection, backface culling, clipping, z-buffering, physics, collisions, and OBJ model import, in about 1,500 lines of Python on top of pygame’s 2D drawing calls.
A palm-sized robot that maps an unknown maze and races back through it. Floodfill pathfinding in under 10 ms per step within 4 KB of RAM on an STM32, with IR sensor fusion and PID wall-following. 1st in division and 4th overall at IEEE MicroMouse @ MIT.
