Dilan Gandhi

ML · Signal Processing · Software
Press any key to begin
SYS BOOT OK — signal locked · 60 BPM
Dilan Gandhi in front of the Manhattan skyline
01 / PLAYER PROFILE

Dilan Gandhi

CLASS Class of 2028 DEGREE B.S. Computer Science · B.S. Finance SCHOOL Rutgers University — New Brunswick HONORS IAFOR Research Symposium (Tokyo) · 1st Place, NJAS · AJAS Lifetime Fellow

I build things at the intersection of machine learning, signal processing, and software engineering — from CUDA-accelerated diffusion pipelines to deep learning models that read ECGs.

CH 01 — Rhythm Runner

· keep the pulse alive

The monitor is live. Hit SPACE (or tap) exactly when the beat crosses the target line. Miss three beats and the patient flatlines. The rhythm speeds up as you go.

SCORE 0 BPM 60 ♥ ♥ ♥

Rhythm Runner

Press SPACE or tap when the spike hits the target zone. Keep the pulse alive.

02 / VITALS

Live Readout

0.0
GPA · 3 national honors
0K+
Lines of code written
0K
Minutes of music / yr
0
Years coding
0+
Research projects
0+
National parks visited
03 / ABOUT

Background

I build things at the intersection of machine learning, signal processing, and software engineering.

I started coding six years ago and never really stopped. What hooked me wasn't apps — it was signals: the idea that a heartbeat, a song, or an image is just a waveform waiting to be understood.

That thread runs through everything I do. At Rutgers I'm pursuing dual degrees in Computer Science and Finance. At the University of Maryland I built CUDA-accelerated pipelines that cut diffusion model training time by 25%. At Rutgers WINLAB I built an 8-bit CPU from breadboards and logic gates. And with CardiacCare, I trained a deep learning model that classifies arrhythmias from raw ECGs at 97% accuracy — work I presented at the IAFOR Undergraduate Research Symposium in Tokyo.

When I'm not writing code I'm usually listening to music — about 80,000 minutes a year of it — or out visiting national parks (15 and counting). Probably why so much of my work ends up being about rhythm, timing, and pulse.

Relevant Coursework

  • Data Structures
  • Algorithms
  • Computer Systems
  • Object-Oriented Programming
  • Linear Algebra
04 / LOADOUT

Skill Board

Confidence Meter

Python██ 95%
Java██ 88%
C / C++██ 85%
CUDA██ 80%
React██ 78%

Also In Inventory

PyTorchNumPyTypeScript SQLKotlinJulia FlaskAWSGCP REST APIsHTML/CSSGit / GitHub Actions LinuxRedisCI/CD · MLOps MATLABR / SPSSFirebase
05 / BUILDS

Projects

06 / CAMPAIGN LOG

Experience

June 2026 — Present

Software Engineering Intern

Amneal Pharmaceuticals

AI-assisted analysis at manufacturing scale — finding failure patterns before the equipment does.

  • Applied AI-assisted analysis across 750K+ sensor process records to identify recurring equipment failure patterns, improving anomaly detection accuracy by 20% across manufacturing and packaging operations
  • Developed data-driven troubleshooting frameworks and standardized diagnostic workflows, reducing manual diagnostic time by 50% and improving equipment reliability across engineering teams
May 2025 — August 2025

Research Intern

University of Maryland — Computer Science Department

CUDA-accelerated diffusion pipelines — making generative models train and run faster.

  • Built CUDA-accelerated training and inference pipelines for diffusion models across 300K+ images — 25% faster training, 35% better GPU throughput, 15% less GPU memory
  • Evaluated diffusion transformer architectures across model scales, benchmarking and optimizing to improve large-scale inference latency by 20%
June 2024 — August 2024

Computer Engineering Intern

Rutgers Wireless Information Network Lab (WINLAB)

Computing from first principles — a CPU built by hand, one logic gate at a time.

  • Built and tested an 8-bit CPU architecture using breadboards, logic gates, and integrated circuits — instruction execution, arithmetic operations, and low-level architecture from scratch
  • Developed automated signal sequencing and validation workflows using clock-cycle analysis to verify instruction execution and control flow across test cases
June 2022 — October 2022

Software Engineering Intern

iStart Valley

First internship, first team lead — shipped a cross-platform mobile app end to end.

  • Led a team of 4–5 interns building Automate, a mobile app for automotive tutorials, maintenance tips, and local meet-ups, using React Native and Firebase
  • Engineered scalable backend systems with Firebase, REST APIs, and cloud functions to automate authentication, content delivery, and event scheduling