Aarya Vulupala Software Engineer
Resume
Menu

Software Engineer Rutgers University CS + Data Science

Aarya Vulupala

I build full-stack products, data pipelines and ML models. Founder of a campus marketplace, developer of the booking platform that runs my own barbershop, and a software engineering intern who shipped to 25+ research offices.

Three things I built, and how they actually work.

Dormly marketplace home page showing a grid of featured listings (shelves, iPhone 15, dumbbells, trampoline, TV, sweater, scooter, pool table) and a sidebar with Jobs, Campus Clubs, Course Exchange, Find a Roommate and Services.
Marketplace home, filtered to Rutgers University

Campus marketplace · Founder

Dormly

Student life is spread across group chats and listing sites. Dormly puts housing, goods, services, jobs, course swaps and clubs in one place, scoped to your campus.

Live. Stripe payments and ML-driven recommendations are next.

listing categories
14
campus hubs
7

What I built

  • Founded and built it end to end: Home, Browse, Listing Detail, Messages and Profile pages in Next.js and React.
  • Designed the PostgreSQL schema behind 14 listing categories and 7 hubs (jobs, roommates, course swaps), with search filtered by campus.
  • Clerk handles sign-up and login with Google and Facebook. Image uploads go to S3, locked to a dedicated folder.
  • Next.js
  • React
  • TypeScript
  • PostgreSQL
  • AWS S3
  • Clerk

How it fits together

  1. Sign in Clerk, via Google or Facebook
  2. Browse Campus-filtered search over 14 categories
  3. List Images stored in a locked S3 folder
  4. Connect In-app messaging between students
Landing page of the Aarya Cuts booking site with the headline "A fresh cut, wherever you are", a Book an appointment button, and a gallery of haircut photos beginning below.
Landing page, with the AI support chat button at bottom right

Booking platform for my barbershop

Aarya Cuts

I run a barbershop, so I built the booking system I wanted: clients pick a time, pay, and get reminders, and the appointment lands on my calendar without me touching it.

Live and in daily use.

appointments / month
60+
monthly revenue
$1.5K+
no-shows
−34%
less scheduling time
~80%

What I built

  • Next.js and React on Vercel. Clients book and pay through Stripe; Supabase (PostgreSQL) tracks appointments.
  • Each booking adds an event to my Google Calendar and notifies me. Clients get a calendar link of their own.
  • Resend sends confirmations and pre-appointment reminders, which cut no-shows by 34%.
  • Support chat on the Gemini API, steered by a custom prompt with the shop’s details so it stays on topic. A short quiz recommends hair care.
  • Next.js
  • React
  • Supabase
  • PostgreSQL
  • Stripe
  • Resend
  • Google Calendar
  • Gemini API
  • Vercel

How it fits together

  1. Book Client picks an open time slot
  2. Pay Stripe checkout
  3. Store Appointment saved in Supabase / PostgreSQL
  4. Sync Event on my Google Calendar, plus a calendar link for the client
  5. Remind Resend emails confirmation and reminder
Project artwork: an NBA-style player silhouette inside a rising bar chart, ringed by the logos of Python, scikit-learn, XGBoost, TensorFlow and SHAP.
Project artwork

Machine learning · Python

NBA Prospect & Career Predictor

Can a draft prospect’s college profile predict the tier of their NBA career? I built the data pipeline and benchmarked models on held-out draft classes to find out.

within-1-tier accuracy (baseline 50%)
79%
prospect records
5,589
data errors
−93%

What I built

  • Python pipeline merging 5,589 prospect records from 4 scraped sources, with automated pandera checks that cut data errors 93%.
  • Benchmarked 7 models, including XGBoost and TensorFlow, across 25 draft classes. The tuned XGBoost reached 79% within-1-tier accuracy against a 50% baseline.
  • Engineered 63 features, lifting macro-F1 57% on held-out draft classes. SHAP showed age, steals and BMI as the top drivers of career tier.
  • Python
  • scikit-learn
  • XGBoost
  • TensorFlow
  • SHAP

How it fits together

  1. Collect 4 scraped sources, 5,589 prospects
  2. Validate pandera checks, −93% errors
  3. Engineer 63 features
  4. Benchmark 7 models over 25 draft classes
  5. Explain SHAP: age, steals, BMI lead

Shipping software, and running things.

Aug 2026 – Present

New Brunswick, NJ

Software Engineer

Rutgers Solar Car Team

  • Built a free sponsor outreach tool in Python and Google Apps Script, using a local LLM to fit a tight budget and a ranking algorithm that prioritizes likely responders based on past sponsors. It emailed 1.5K+ sponsors and grew funding from $100 to $2,000+.
  • Built a QR code check-in tool: members scan to update a live attendance dashboard that reports to the team president automatically.
  • Analyzed 22 telemetry channels (speed, RPM, drag, GPS) and 20 rival-team datasets with NumPy, pandas and Databricks to run physics-based race simulations with anomaly alerts, boosting simulated race speed by 14%.
The Rutgers Solar Car team posing beside the solar car outside a building.
  • Python
  • Apps Script
  • NumPy
  • pandas
  • Databricks

Jun 2026 – Aug 2026

Remote

Software Engineer Intern

IpserLab

  • Shipped 15+ full-stack features in Java, React and PostgreSQL for a tech-transfer platform used by 25+ research offices, including REST API endpoints and database changes. Also fixed front-end bugs and added safety checks that block prompt-injection attacks on the AI chatbot.
  • Led the move from JavaMail to the Brevo API for ~5K monthly emails after diagnosing why Microsoft-hosted inboxes were rejecting messages. Bounce rates fell 60%.
  • Built notification settings in React on top of Slack, Twilio and Brevo APIs, so users choose email, text or Slack alerts in one place.
Tectra Labs home page with the headline "From lab to launch — fueling innovation, building futures."
  • Java
  • React
  • PostgreSQL
  • REST APIs
  • Brevo
  • Twilio
  • Slack

Jun 2025 – Aug 2025

Plainsboro, NJ

Freelance Software Engineer

Aljon’s Grill Pizzeria & Restaurant

  • Built a Python/SQL pipeline that consolidated 5K+ orders from 5 sales channels (DoorDash, online, phone, in-person) into one PostgreSQL database, cutting weekly reporting from 3 hours to 10 minutes.
  • Developed a Next.js/React app where 15+ non-technical kitchen staff log stock and waste daily across 50+ ingredients.
  • Added sales and inventory views that showed the owner which items were over-ordered, cutting food waste by roughly $30 a day.
  • Python
  • SQL
  • PostgreSQL
  • Next.js
  • React
  • HTML
  • CSS

Leadership

Sep 2026 – Present

New Brunswick, NJ

Software Engineering Fellowship Education Director

Blueprint

  • Instructed 100+ students, designing a full curriculum on system design, data structures, algorithms, HTTP, REST and runtime complexity.
  • Wrote hands-on labs in Python, FastAPI and SQL, mentoring fellows on REST API design, login security, database modeling and scalability.
  • Python
  • FastAPI
  • SQL

Apr 2026 – Present

New Brunswick, NJ

President

Iota Nu Delta Fraternity, Inc.

A multicultural, South Asian interest fraternity

  • Led the nation’s largest Iota Nu Delta chapter, driving 33% membership growth and allocating a $25,000 budget across 12 roles twice a year.
  • Directed weekly chapter meetings for 40 members, setting short- and long-term priorities and delegating work across 12 officer positions.

Rutgers University

Computer Science & Data Science

Relevant coursework

  • Computer Architecture
  • Data Structures and Algorithms
  • Computing in Java
  • Database Management
  • Discrete Structures
  • Linear Algebra
  • Calculus 1–3
  • Statistics

What I reach for.

Languages

  • Java
  • Python
  • JavaScript
  • TypeScript
  • SQL
  • C
  • HTML
  • CSS

Libraries & frameworks

  • React
  • Next.js
  • Node.js
  • Spring Boot
  • FastAPI
  • pandas
  • NumPy
  • scikit-learn
  • XGBoost
  • TensorFlow

Cloud & databases

  • AWS (EC2, S3, RDS, Lambda)
  • PostgreSQL
  • MySQL
  • Firebase
  • Supabase
  • REST APIs

Tools

  • Git
  • GitHub
  • Databricks
  • CI/CD
  • Postman
  • Claude Code
  • Codex
  • Cursor

Hiring, collaborating, or just curious about something I built? Email is fastest.

↑↓ move↵ runesc close