Adharsh Jaikumar

Adharsh
Jaikumar

Computer Engineering — Virginia Tech

Connect Follow

01 — Selected Work

Research

01EyeNova 02ALASCA

02 — Open Source

GitHub

@adharshjai ↗

03 — Curriculum Vitae

Resume

View Full Resume ↗

Experience

Hybrid Optical-Electrical Nano-Bio Interface Lab

Research Intern, Virginia Tech Programmable Cloud Laboratory | Blacksburg, VA May 2026 – Present

Developed AI-enabled software to optimize "Design-Build-Test-Learn" workflows for semiconductor fabrication systems.
Collaborated with researchers and engineers to develop AI-driven solutions for next-generation healthcare and semiconductor technologies.
Selected for the Virginia Tech Programmable Cloud Laboratory node of a $20M NSF-funded initiative focused on ML-enabled real-time biosensing.

Louisville Automation & Robotics Research Institute

Student Research Intern, University of Louisville | Louisville, KY May 2025 – Aug 2025

Developed machine-learning based technology to identify and reconstruct robotic movements using audio recordings, revealing potential security risks in industrial automation.
Co-authored research on acoustic side-channel attacks accepted at IROS 2026; earned 2nd Place at the UofL CSE Graduate Research Competition.

Education

Virginia Tech

Bachelor of Computer Engineering Expected May 2029

GPA: 4.0/4.0 · Richard H. Warren Scholar | Blacksburg, VA

Research

EyeNova: Ophthalmic Imaging for Early Disease Detection

Python, PyTorch, OpenCV, Slit-Lamp Optics

Built a low-cost slit-lamp imaging system using commodity optics and developed a CNN pipeline to detect ocular biomarkers associated with Wilson's disease, glaucoma, diabetic retinopathy, and macular degeneration.
Trained and evaluated image-classification models achieving 93.43% accuracy for Kayser–Fleischer ring detection and 89.40% accuracy for glaucoma detection.
Won 1st Place at a regional science fair, 2nd Place at the Kentucky State Science Fair in Computational Biology, and multiple Air Force Research Laboratory Awards.

Projects

Myelin: Supervised Autonomous Browser Agent

TypeScript, Electron, Playwright/CDP, Claude API

Built an 8,000-line, 85-module TypeScript desktop agent that analyzes arbitrary web forms, maps requested information to a structured user profile, and completes fields through embedded Chromium without per-action user approval.
Developed an LLM-based form-understanding layer that maps previously unseen DOM structures onto typed profile fields while blocking unsupported or unsourced claims before submission.
Implemented typed Electron IPC, local-only SQLite storage, and dedicated Greenhouse, Lever, and Ashby adapters to support common application workflows across different site structures.

Trauma Relief Rover

Jetson Xavier NX, CUDA, Robot Operating System (ROS), mmWave Radar

Built a ROS-based perception system on a Jetson Xavier NX that processes mmWave radar data to detect human presence and subtle motion through visually obstructed environments for search-and-rescue applications.
Moved range-Doppler signal processing onto the GPU using CUDA to support real-time radar analysis and integrated resulting detections into the rover's ROS perception pipeline.

Stock Trading Platform

Python, Streamlit, scikit-learn

Built a Streamlit application that transforms historical market data into technical-indicator features, applies scikit-learn models to predict price trends, and backtests model-driven decisions against a buy-and-hold baseline.

Publications

[1] Ayan Joshi, Adharsh Jaikumar, Saveer Jain, Anomadarshi Barua, Sabur Baidya. "ALASCA: Asset Leakage with Acoustic Side-Channel Attacks in Multi-Joint Collaborative Robots." Accepted to IEEE/RSJ IROS 2026.

Skills

Languages

Python Java C++ C# TypeScript HTML/CSS SQL

Frameworks

React TensorFlow PyTorch OpenCV scikit-learn MATLAB

Technical

AWS Google Cloud Firebase Fusion 360 Arduino ROS Docker Android Studio NumPy Signal processing

Certifications

IBM Machine Learning with Python (Honors) IBM Computer Vision & Image Processing