0%
click to skip
// system.init — open to internships & full-time roles · CPT Eligible

Onasvee
Banarse.

MS CS @ UMass Amherst · |

Engineer who has spent several years breaking systems in order to design and deliver new systems through Reliance Jio. Presently working towards a Master's degree in Computer Science, focusing on Machine Learning, Software Engineering and Quality Assurance, to create fast, reliable systems.

// press [ ` ] or click >_ to launch terminal

View My Work Get in Touch View Resume
Onasvee Banarse
UMass · MS CS
scroll to explore
// about.me

The Story
So Far.

I'm Onasvee Banarse, a graduate CS student at UMass Amherst, and a published researcher with a background in QA automation and cybersecurity.

Before grad school, I was an SDET / Associate QA Engineer at Reliance Jio Platforms — one of India's largest tech companies where I built Selenium frameworks, validated OAuth2/OIDC SSO flows, and cut regression effort by 10+ hours a week.

My research spans deep learning for image forensics and GANs, with three co-authored publications. I bring both engineering rigor and research curiosity to everything I build.

// Logistical Status
Based in Amherst, MA · Actively seeking Fall Internships in the US · CPT Eligible · Available for on-site and remote roles.

2+
Years Industry QA Exp
9.38
B.Tech GPA / 10
3
Research Publications
3.433
Master's GPA / 4
// education

Where I've
Studied.

Master of Science — Computer Science
// University of Massachusetts Amherst
// relevant coursework Machine Learning · Advanced Algorithms · Distributed Systems and Operating Systems · Information Retrieval · Neural Networks: Modern Intro · Data Science Fundamentals
Jan 2026 — Dec 2027
In Progress (Current GPA : 3.43/4)
B.Tech — Computer Science
// Savitribai Phule Pune University, India
// relevant coursework Programming & Problem Solving · Object Oriented Programming(OOP) · Principles of Programming Languages · Object oriented Modeling & Design · High Performance Computing · Data Structures & Algorithms · Design & Analysis of Algorithms · Systems Programming & Operating System · Database Management Systems · Computer Networks & Security · Distributed Systems · Data Science & Big Data Analytics
Aug 2019 — Mar 2023
GPA 9.38 / 10
// skills.stack

What I
Work With.

// languages
Python Java JavaScript C/C++ SQL (PostgreSQL) Groovy HTML/CSS
// frameworks & tools
Selenium JMeter SoapUI Node.js Express.js Postman Docker Git
// ml / data science
TensorFlow Scikit-learn NumPy pandas Matplotlib ResNet CNNs
// databases
PostgreSQL CassandraDB MongoDB
// work.history

Where I've
Been Before.

Dec 2023 — Dec 2025
Navi Mumbai, India

SDET / Associate QA Engineer

// Reliance Jio Platforms Limited
  • Built Selenium automation frameworks in Java for web testing, integrating API calls to validate backend data against UI — saving 10+ hours/week on regression.
  • Achieved 85% test coverage on critical IDAM modules via JMeter, Postman, and SoapUI load & functional testing.
  • Validated JioID SSO end-to-end: OAuth2/OIDC flows, token handling, and multi-client configurations.
  • Reduced defect leakage by 30% through cross-database validation (JMeter JDBC + SQL) across Cassandra and PostgreSQL.
SeleniumJava JMeterSoapUI OAuth2/OIDCPostgreSQL CassandraDB
Jan 2022 — Jun 2022
Bengaluru, India

Cyber Security Intern

// Acmegrade Pvt Ltd
  • Conducted OWASP-methodology vulnerability assessments and penetration testing on internal web applications.
  • Reduced false positives by 60% through Wireshark and SIEM alert tuning and traffic analysis.
  • Reverse-engineered malware samples to build detection signatures for internal threat intelligence systems.
OWASPPenetration Testing WiresharkSIEM Malware Analysis
Apr 2021 — Nov 2021
India

Data Science Intern

// RegEx Software Services
  • Conducted EDA on the Kaggle Credit Card Fraud Detection dataset (284,807 records), improving feature selection accuracy by 2.62%.
  • Built data visualizations with Matplotlib and Seaborn that drove key product recommendations for the risk analysis team's fraud detection thresholds.
  • Assisted ML model deployment on AWS EC2/S3 using Docker and Flask — cutting deployment time by 50%.
  • Final project: Credit Card Fraud Detection model achieving 96% accuracy and 93% reduction in false positives.
PythonScikit-learn MatplotlibSeaborn AWS EC2/S3DockerFlask
// selected.work

Things I've
Built.

// project_06 · Jul 2026

F1 Pit Stop Strategy Engine

Real-time recommendation engine and Streamlit dashboard that calculates optimal Formula 1 pit stop windows. Leverages Monte Carlo simulations and XGBoost regression trained on historical FastF1 lap data.

✦ GPU-accelerated real-time backtesting capabilities
// my role Architected the machine learning pipeline, integrated PyTorch CUDA acceleration for simulation processing, and designed the interactive dashboard.
PythonXGBoost PyTorch CUDAStreamlit SQLiteFastF1
GitHub
// project_07 · May 2026 – Jul 2026

Local AI Audiobook Reader

Offline audiobook pipeline for manga and light novels that performs character voice attribution and visual text synchronization. Extracts PDF text and utilizes local LLMs paired with Edge TTS for expressive character dialogue.

// my role Built the end-to-end system combining EasyOCR, text parsing, and local Llama 3.1 8B integration via Ollama to run entirely offline.
PythonLlama 3.1 8B OllamaFlask EasyOCREdge TTS
GitHub
// project_01 · Jul 2022 – Jul 2023 · Team of 4

Image Manipulation & Forgery Detection

CNN-based deep learning pipeline to detect manipulated regions in digital images. Trained on CASIA v1.0 and v2.0 datasets, covering copy-move, splicing, and retouching. ELA used for feature extraction; deployed with a Streamlit interface for real-time inference. Results consistent with 20+ state-of-the-art papers reviewed.

✦ 95.30% accuracy on CASIA v1.0 · 91.31% on v2.0
// my role Designed and implemented the ResNet-based image classification pipeline for detecting manipulated regions, optimising performance across diverse manipulation types.
PythonTensorFlow ResNetUNet StreamlitELA
GitHub Published Paper
// project_02 · Aug 2024 – Nov 2025 · Reliance Jio

Data Migration Utilities

REST-based migration services to move legacy CassandraDB records into restructured schemas following table field changes. Atomic PostgreSQL transactions + idempotent Cassandra writes ensure reliability. Validated with pre/post-migration checksums — zero data loss confirmed post-deployment.

✦ Adopted by the backend engineering team for production use
// my role · solo ownership Solely designed, built, and deployed the entire migration utility. Owned the full lifecycle — architecture, implementation, validation, and handoff to the backend engineering team.
Node.jsExpress.js PostgreSQLCassandraDB REST API
Private · Built at Reliance Jio
// project_03 · Jan 2025 – Nov 2025 · Reliance Jio

Test Automation Management Tool

Personal POC that grew into a team-wide tool where it automatically organising fragmented JMeter and SoapUI test suites by module and priority tags. Groovy scripts dynamically parse test files and generate regression-ready configurations for repeatable multi-environment execution.

✦ Saved 1–2 working days per release cycle · Adopted by the QA team
// my role · initiated & led Built the Python organiser tool independently. Also championed the migration from JMeter to SoapUI for API testing — identified the opportunity, ran the POC, and introduced it to the team (concept suggested by manager). Implemented all Groovy scripting end-to-end.
PythonJMeter SoapUIGroovy
Private · Built at Reliance Jio

More in the works...

I am constantly building, breaking, and experimenting with new systems. Catch my latest commits and ongoing work over on GitHub.

View GitHub Profile
// research.published

Published
Research.

Unmasking the Deception: An Image Manipulation Detection Using ResNet and UNet Architecture
Onasvee Banarse, Harsh Shah, Kaustubh Kabra, Akash Mete
International Research Journal of Engineering and Technology (IRJET) · Vol. 12, Issue 03 · Mar 2025
// tl;dr summary Developed a deep learning framework achieving 95.3% accuracy on CASIA datasets by pairing Error Level Analysis (ELA) with ResNet/UNet architectures to detect localized image splicing, copy-move, and retouching manipulation regions.
2025
01
Exploring the Potential of Generative Adversarial Networks: A Comparative Study of GAN
Harsh Shah, Kaustubh Kabra, Onasvee Banarse, Akash Mete
International Research Journal of Engineering and Technology (IRJET) · Vol. 10, Issue 03 · May 2023
// tl;dr summary Conducted a technical comparative analysis of generative model architectures (Vanilla GANs, DCGANs, and CycleGANs), evaluating structural training stability, loss function convergence profiles, and image synthesis quality.
2023
02
Autonomous UAV Swarms: Powering Up with AI
Onasvee Banarse, Harsh Shah, Kaustubh Kabra
International Journal for Research in Applied Science & Engineering Technology (IJRASET) · Vol. 11, Issue 06 · Jun 2023
// tl;dr summary Investigated decentralized multi-agent control systems for unmanned aerial vehicle (UAV) swarms, analyzing edge-computed path planning algorithms and collision-avoidance mechanisms for collaborative civilian monitoring operations.
2023
03
// certifications.leadership

Awards, Certs
& Leadership.

// awards & certifications
⭐
Star Performer Award
Reliance Jio Platforms · Jul 2025
Awarded by Mr. Anish Shah, President of JIO Platform Limited.
🧠
Neural Networks and Deep Learning
DeepLearning.AI / Coursera · Jun 2023
☕
AI & ML in Java · Java Programming · Java Foundations · PL/SQL
Oracle Academy · 2020 – 2022
📡
AI/ML Fundamentals · Linux & Data Analytics
Vodafone India · Jan – Mar 2021
// leadership & activities
AI ML Lead — CodeHub (Coding Club)
AISSMS Institute of Information Technology · 2022 – 2023
Led the AI/ML vertical of the college coding club, organising workshops and technical sessions.
AIML Co-Lead — Microsoft Learn Student Ambassadors
MLSAC, AISSMS · 2021 – 2023
Co-led the AI/ML chapter under Microsoft's student ambassador programme for two years.
Technical Secretary & Technical Representative
Computer Engineering Students' Association · 2021 – 2023
Served as Technical Representative (2021–22) then Technical Secretary (2022–23), coordinating departmental tech events.
// contact.init

Let's
Work Together.

I'm actively looking for internship and full-time opportunities in software engineering, ML, or QA automation. My inbox is open — let's talk.

Email LinkedIn GitHub Resume +1 413-275-8451
// journal Paper Title
Open in Tab
loading via google docs viewer...
📄
// viewer unavailable Google Docs viewer couldn't load this PDF. Click below to open it directly.
Open Paper
// easter_egg
// CLASSIFIED — LEVEL 5 CLEARANCE REQUIRED
CONFIDENTIAL
Subject Dossier // Asset #22-OB
Subject Name
Onasvee Banarse
Current Coordinates
Amherst, MA [ACTIVE]
Primary Objective
Secure Fall Internship in Software Engineering / AI & ML
Known Aliases
ORION-22, CLASSIFIED
Past Affiliations
Reliance Jio Platforms [2023—2025]
Status: Mission Accomplished
Skills & Arsenal
Weapons of Choice
Python, Java, Selenium, TensorFlow, PostgreSQL
Special Abilities
Reducing regression effort by 10+ hours/week. Image Forgery Detection. Breaking APIs to build better ones.
// END OF FILE. THIS MESSAGE WILL NOT SELF-DESTRUCT.