Shaunak Badani

Shaunak Badani

M. Eng in AI @ Duke

Check out my blog!

About Me

I am an AI grad student at Duke, with 2 years of experience in Software Engineering at Oracle, and 3 years of undergraduate research experience in Computational Natural Science at IIIT Hyderabad. I excel at learning new concepts fast, and I love to think about complex problems from different perspectives.

Experience

Software Engineer Intern

Oracle NetSuite

May 2025 - August 2025

Redwood City, CA

  • Improved a tool for Technical Support staff that aids them in resolving customer issues without having to transfer it to the Engineering team, resulting in speedier resolution.
  • Added the `ConsignmentLink` table in the system, which was needed as part of a larger feature `Vendor Consignment`.
  • Resolved multiple customer issues, one of them impacting the workflow of 7 customers.

DevOps Engineer Intern

Quizzi.AI

Jan 2025 - Feb 2025

Durham, NC

  • Implemented hotswap for the docker development setup, which enables developers to see changes to the application instantly, decreasing development time by 60%.
  • Reduced final build images by changing base image from python to it's alpine in the backend, and only copying build files to the image in frontend, reducing the image size by 97.53 %.
  • Removed environment files from docker compose files and added .env files, thereby reducing the risk of API key leakage.
  • Implemented session based authentication, which persisted login states after browser refreshes.
  • Converted front end pages from stateful pages to stateless, which allowed users to refresh at different states of the application without losing context.

Software Engineer

Oracle Netsuite

June 2022 - July 2024

Hyderabad, India

  • Developed a feature that allows users to create reversals of Journal Entries using two clicks, which otherwise needed manual input.
  • Enhanced developer tools to assist in identifying common customer issues with the Currency Revaluation engine, which saved several weeks of manual developer investigation and resulted in faster issue resolution.
  • Added slack notifications for team-city (CI / CD) failures that help detect build failures early on during feature development, which saved precious time during feature integration into production.
  • Assisted in transitioning the team from Perforce to Git as the VCS agent, ensuring a smooth transition of workflow well before support for Perforce from the Dev-Ops team was discontinued.
  • Held informational sessions on several topics - on both domain and technical knowledge, which helped the team gather baseline knowledge necessary at work.
  • Resolved multiple high-urgency customer issues within 1-2 business days by investigating and checking product expectation from documentation.
  • Integrated the code-fix and automated test of a product issue by making code changes that was blocking a work-flow for 38 customers.
  • Mentored new-hires in product development and domain-related knowledge, with a goal to gradually make them independent in their work.

Web Administrator

IIIT Hyderabad

August 2019 - October 2020

Hyderabad, India

  • Added a field of ”year” on the course-feedback portal by making changes to the code-base, that helped faculty separate 2020 and 2021 course feedback data seamlessly.
  • Deployed a newer version of the ”Central Authentication Service” (CAS) by modifying the template, which allowed integration of other services with CAS that were non-compliant with older versions of the latter.

Research student

IIIT Hyderabad

July 2019 - October 2022

Hyderabad, India

  • Implemented an enhanced molecular dynamics algorithm in python and C++ to achieve simulation convergence quicker.
  • Published a research paper outlining the derivation of the algorithm and its implementation on three simple systems and one molecular system
  • Organized informational sessions on foundational statistical mechanics concepts that helped juniors in the research group graduate faster.

Software Developer Intern

Kidsens

May 2019 - August 2019

Hyderabad, India

  • Implemented a mobile application in React Native that presents students with a set of questions that helps evaluate whether a kid at school needs a medical check-up
  • Mocked up application UI’s in Photoshop before implementing the UI, which helped the team arrive at a consensus on a good UI design.
  • Developed the backend API in django and deployed the application on an AWS EC2 instance.

Education

M.Eng in Artificial Intelligence

Duke University

August 2024 - December 2025

Durham, NC

Focus: Explainable AI for diffusion models

M.S in Computational Natural Sciences

International Institute of Information Technology, Hyderabad

August 2021 - November 2022

Hyderabad, India

Focus: Algorithmic implementation of a novel molecular dynamics method

B.Tech in Computer Science

International Institute of Information Technology, Hyderabad

August 2017 - July 2021

Hyderabad, India

Focus: Web development, Operating systems, Optimization methods, Statistical methods in Artificial Intelligence

Technical Expertise

AI/ML

  • PyTorch
  • TensorFlow
  • Autograd

Programming

  • Python
  • JavaScript
  • C++
  • CUDA
  • Java

Cloud & MLOps

  • Azure App Services
  • Nginx
  • Docker

Web Development

  • React
  • Flask
  • FastAPI
  • Django

Data Science

  • Pandas
  • NumPy
  • Scikit-learn
  • Matplotlib
  • PyTorch

Tools

  • Git
  • Linux
  • Bash scripting

Featured Projects

WordWeight

WordWeight lets you generate stunning images from text using cutting-edge diffusion models — then pulls back the curtain to reveal which words carried the most weight.

PyTorch
Diffusion models
Replicate
PostgresQL
Devil's Codex

A chatbot application that assists in answering questions about the Duke AI MEng Program, events around Duke, and questions for prospective students.

OpenAI function calling
Agentic Framework
FastAPI
React
Docker
Macrograd

Built a custom machine learning framework (like Pytorch) in C++

Python
Numpy
C++
Pytorch
CMake
GMock
Replica Exchange with Non-equilibrium switches

Implemented an enhanced molecular dynamics algorithm that helps speed up simulation convergence by employing non-equilibrium methods of simulation.

Python
C++
Numpy
Matplotlib

Get in Touch