Tushar Mhatre | Data Scientist

Projects

Flappy Kiro: Cloud Gaming Platform

An enterprise-grade cloud gaming platform demonstrating modern DevOps, Kubernetes orchestration, and comprehensive observability on AWS. Features 24+ pods across microservices, sub-100ms response times, 99.9% uptime, and a full monitoring stack with distributed tracing and real-time alerting.

Kubernetes AWS (EKS, ECR, ALB) Docker Python Flask Nginx OpenTelemetry Infrastructure as Code DevOps

GitHub

Data Analysis Agent

An AI-powered Exploratory Data Analysis agent. Upload your dataset, ask questions in plain English, and get instant insights with auto-generated Python code and visualizations. Features LangGraph orchestration, secure sandbox execution, and rich visualizations using Matplotlib, Seaborn, and Plotly.

React TypeScript FastAPI LangGraph LangChain Python Groq API Docker Pandas Data Visualization

GitHub Live Demo

AI Research Paper Assistant

An offline-capable RAG-powered chatbot for secure, efficient summarization and Q&A on academic papers.

Langchain Streamlit Large Language Models Ollama API FAISS Sentence-Transformers (MiniLM)

GitHub

Fake Job Detection Portal

Leveraged NLP and ML to detect fraudulent job postings, using classification models and a Flask-based web app for real-time predictions.

Python Flask Natural Language Processing HTML/CSS Statistical Analysis Smote

GitHub

Predicting Customer Revenue

Implemented Linear Regression, Lasso, Ridge, and M.A.R.S. using R (caret) to forecast customer revenue, applying 10-fold cross-validation.

R Caret Regression Statistical Analysis

GitHub

Hospital Readmission Prediction

Trained Logistic Regression, Random Forests, and XGBoost models to predict patient readmission likelihood based on medical histories.

R Programming Caret GGplot Classification

GitHub

Database Design for PAN

Built an Azure SQL database and Java/JDBC application for patient assistance, including ER diagram, schema design, and indexing.

Java Azure SQL Database Design JDBC Azure Data Studio

GitHub

Financial Analysis of a Supermarket Chain

Created interactive Tableau dashboards to identify key products and demographics to optimize marketing and sales strategies.

Tableau Data Visualization

Tableau

Predicting Hospital Length of Stay

This project aims to analyze medical data to predict the length of stay (LOS) of patients in a hospital. Various machine learning models are trained and evaluated to predict the LOS based on demographic and medical information.

Python Pandas Seaborn Matplotlib Scikit-learn

GitHub

Predicting Smoking & Drinking Habits with EHR Data

An analytics pipeline developed in R that predicts individuals’ smoking and drinking behaviors using a suite of classification models

R Tidyverse ggplot2 caret stats

GitHub

Professional Experience

September 2025 - Present

AI Engineering Intern

The Aether Loop, Remote

Architected and built an Agentic AI recommendation engine by leveraging CrewAI, LangChain, and structured clinical insight extraction with SMEs, achieving NDCG of 0.91.
Improved model performance from NDCG 0.60 to 0.91 through continuous feedback and human-in-the-loop validation, reducing SME review time from 4 to 1 hour weekly, leading to faster iteration cycles.
Engineered and maintained a containerized FastAPI backend for deployment cycles, sustaining 99% uptime with reliable monitoring and continuous model improvements, by deploying with Docker on GCP Cloud Run and automated CI/CD pipelines.
Researched and integrated open source tools and third-party APIs, eliminating the need to build several modules from scratch and reducing development time of our mobile app and platform services by ~40–50%.

July 2025 - September 2025

Data Architect Intern

Optimal Living Systems, Remote

Designed and deployed the data ingestion and ETL pipelines in Airflow, OCR (Docling) and OpenAI API, to transform 1,000+ research PDFs into structured QnA datasets to kick off the organization's first LLM fine-tuning pilot.
Built reusable Python (BeautifulSoup, Requests) scraping scripts for the research team, reducing their manual sourcing by 3-4 hours daily, leading to completion of data targets two weeks ahead of deadline.

February 2024 - May 2025

Graduate Research Assistant

University of Oklahoma

Built executable EEG/MRI BIDS transformation pipelines in MATLAB, reducing manual structuring time from 5+ hours per dataset to single-click execution which enabled background preprocessing while researchers prioritized core research work.
Developed WSL-based Bash scripts integrating Windows/MATLAB workflows with Linux tools (FSL, AFNI), reducing preprocessing runtime by 40% and accelerating experiments turnaround through parallel job scheduling.
Led a year-long neuroimaging research study applying clustering algorithms across EEG datasets to identify and analyze Cortical Activity Patterns (CAPs), validating findings through statistical testing and co-authored a manuscript under review at Brain Topography.

February 2021 - July 2023

Software Engineer

Tata Consultancy Services, Pune, India

Developed 9 RPA bots using Automation 360, VBA, and SQL over 6 months, reducing manual workload by 20-25 hours weekly and driving $80K+ annual cost savings for enterprise clients through SAP and mainframe automation.
Designed and deployed ETL pipelines (Airflow) to aggregate data from multiple sources (SAP, mainframes) for forecasting models.
Diagnosed and resolved 30+ production incidents in live forecasting pipelines (Python, Airflow), implementing fixes that reduced error rates by 30% and improved system uptime significantly.
Built automated reporting solutions using Power Query, Excel automations, and multi-source connectors, eliminating 10+ manual hours weekly and enabling consistent daily-refresh data pipelines.
Developed and containerized ML models using Docker and exposed them through lightweight Flask APIs, reducing latency from 2s to 300ms while handling 1K+ RPM.
Collaborated with cross-functional teams in Agile sprints to gather requirements, define project scope, and deliver data solutions aligned with business needs.

Education

Master's

August 2023 - May 2025

University of Oklahoma

Data Science

GPA: 3.97

A hands-on Curriculum Focusing on underlying concepts of Data analytics, Machine learning and Deep Learning

Relevant Courses

Data Mining
Machine Learning Practice
Intelligent Data Analytics
Data Visualization
Statistics

Achievements/Extra-Curriculars

Managed to get Graduate Research Assistantship
Data Science Club - Member

Bachelor's

June 2016 - May 2020

Savitribai Phule Pune University

Electronics & Telecommunication Engineering

GPA: 7.6

Gained a robust foundation in object oriented programming, data structures, mathematics and technical problem-solving.

Relevant Courses

Fundamentals of Programming-I and II
Data Structures and Algorithms
Object Oriented Programming
Engineering mathematics-I,II & III

Achievements

Telecom Engineer Student Assosciation - Treasurer(2019)
Telecom Engineer Student Assosciation - Assistant Designer(2017)
Rotaract Club of DPU - Assistant Marketing Head(2017)

Hi, I'm Tushar Mhatre

Welcome to my PortFolio

Projects

Flappy Kiro: Cloud Gaming Platform

Data Analysis Agent

AI Research Paper Assistant

Fake Job Detection Portal

Predicting Customer Revenue

Hospital Readmission Prediction

Database Design for PAN

Financial Analysis of a Supermarket Chain

Predicting Hospital Length of Stay

Predicting Smoking & Drinking Habits with EHR Data

Professional Experience

AI Engineering Intern

Data Architect Intern

Graduate Research Assistant

Software Engineer

Education

Relevant Courses

Achievements/Extra-Curriculars

Relevant Courses

Achievements

Skills

Programming Languages

Data Analysis Tools/Skills

Python Libraries

R Packages

Databases

Certifications

Contact Me

Let's work together!

Phone

Email

LinkedIn