0%
Rishikesh Donthula

Data Scientist & Analytics Engineer

Rishikesh
Donthula

MS Analytics @ Georgia Tech · ML · Statistical Modeling · Data Engineering

View My Work
Open to Full-Time & Internship Opportunities

🇺🇸🇮🇳

4

Projects ↗

280K+

Transactions ↗

~85%

ROC-AUC ↗

Open

to Opportunities

About Me

Who I Am

I build ML systems end-to-end, from raw data to working model. My first year at Georgia Tech's MS Analytics program produced three deployed projects spanning churn prediction, computer vision, and unsupervised fraud detection. I'm carrying a 4.0.

My background combines computer science, mathematics, and finance. I can write the query, build the model, and understand what the output actually means for a business.

That combination doesn't come standard.

Outside of work, I follow financial markets closely and spend time on ML research. I tend to gravitate toward problems where the math is interesting and the stakes are real.

Graduate

Master of Science in Analytics

Aug 2025 – Dec 2026

Georgia Tech

Georgia Institute of Technology

Atlanta, GA

GPA 4.0

Coursework

Machine Learning Computer Vision Data Analytics in Business Understanding Markets with Data Science

Undergraduate

Bachelor of Science in Computer Science

Sep 2019 – May 2024

NYU

New York University

New York City, NY

Dean's List · Math & Finance

Coursework

Artificial Intelligence Data Science Databases Data Analysis Multivariate Calculus

Technical Toolkit

Skills

Languages

Python R SQL C++ JavaScript HTML LaTeX

Libraries & Frameworks

PyTorch scikit-learn pandas NumPy SciPy Matplotlib seaborn OpenCV CatBoost XGBoost LightGBM Flask React Native Expo D3.js

Tools & Platforms

Jupyter Git GitHub Hadoop Apache Spark AWS GCP Azure PostgreSQL SQLite OpenRefine MS Project

Methods & Certs

PCA / SVD SVM Random Forest Transfer Learning RANSAC Transformers NeRF SMOTE TF-IDF / LSI PageRank Time Series Clustering JWT Auth Michigan Applied Data Science U. of Michigan

Portfolio

Projects

Churn Prediction

November 2025 · Machine Learning

Customer Churn Prediction & Behavioral Segmentation

~85% ROC-AUC

End-to-end churn prediction across 7 telecom datasets (7,000+ customers). Top 10% highest-risk customers accounted for ~77% of churn. Evaluated 28 configurations; CatBoost + SMOTE was the top performer.

→ Lets telcos concentrate retention spend on the accounts most likely to leave — before they do.

Python CatBoost XGBoost scikit-learn pandas
Semantic Segmentation

October 2025 · Computer Vision

Semantic Segmentation for Urban Scene Understanding

~93% mIoU · Road Seg

Pixel-level labeling of roads, vehicles, and pedestrians in urban driving scenes. ~62% mIoU on an 11-class dataset; ~93% mIoU on binary road segmentation via transfer learning. Compared baselines against PSPNet with dilated convolutions.

→ Foundational perception layer for autonomous vehicles, robotics, and smart city infrastructure.

PyTorch OpenCV PSPNet Python NumPy
Fraud Detection

September 2025 · Unsupervised Learning

Fraud Detection via Unsupervised Pattern Discovery

3× over baseline

Unsupervised fraud detection on 280,000+ financial transactions with no labeled data. Well-separated behavioral clusters with up to 3× improvement over random baselines. K-Means and GMM evaluated via Silhouette Score and Adjusted Rand Index.

→ Works without labeled fraud data — deployable in any financial system from day one.

Python K-Means GMM scikit-learn Matplotlib

TradeEasy

P2P Barter · iOS

May 2023 · Full-Stack Mobile

TradeEasy — Peer-to-Peer Bartering App

End-to-End · iOS · AWS

Mobile marketplace for item trading using a Tinder-style swipe interface. Flask + PostgreSQL backend on AWS RDS with S3 image storage. Location-aware matching via geodesic filtering, JWT auth, and real-time chat.

→ Extensible to any peer-to-peer exchange market — goods, services, or skills.

React Native Flask PostgreSQL AWS Expo JWT

Work History

Experience

Frontend Development Intern

June 2023 – August 2023

GoFloaters↗

Chennai, India

React Native · JavaScript · Git

  • Improved UI consistency across core screens — Home, Search, Profile, Space Details — on both mobile and web by implementing design spec updates and resolving layout issues.
  • Implemented client-side search logic including filtering and sorting rules to surface relevant inventory in the expected order.
  • Fixed data rendering edge cases in listings and detail views, ensuring pricing and ratings displayed correctly across all space types.

Get In Touch

Let's Chat

Every interesting project starts with a conversation. What are you working on?

Summer 2026 Internships Full-time · Starting Dec 2026 🇺🇸 US Citizen · No sponsorship required

Full CV

Résumé

One page. Everything that matters.

View Résumé↗

It always seems impossible until it’s done.

— Nelson Mandela