Om Kherde

Neuroscience and brain-computer interfaces.

Sophomore at UT Austin studying neuroscience, with certificates in computer science and brain-computer interfaces. I work in the Lega Lab at UT Southwestern on a real-time intracranial EEG pipeline for closed-loop neuromodulation. Before that, drug discovery at Mayo Clinic and EEG classification at SMU.

Email · Google Scholar↗ · LinkedIn↗ · Instagram↗ · CV

Research

Lega Lab, UT Southwestern
Researcher, 2026 Real-time Python/Simulink pipeline streaming 256-channel intracranial EEG at 30 kHz for closed-loop neuromodulation.
Longhorn Neurotech
Researcher, 2025– Motor imagery classification from OpenBCI EEG using CSP and Kalman filtering.
Virtual Drug Screening Lab, UT Austin
Researcher, 2025– Structure-based virtual screening of small-molecule inhibitors, with wet-lab validation.
SMU Governor’s Science & Technology Academy
Invitee, 2025 Single-channel EEG classification of cognitive state. Two ML frameworks for separating mental states.
Mayo Clinic, Caulfield Lab
Student Intern, 2024–25 In-silico validation of an Alzheimer’s target; deep learning models for candidate selection.
St. Jude Children’s Research Hospital
Mentee, 2024 Attention-based networks for target identification.

Publications

Evaluating the Effectiveness of Suv39h1: A Novel Epigenetic Factor in Alzheimer’s Disease as a Therapeutic Target in Drug Discovery. Academia, 2024.

All work on Google Scholar↗

Independent work

Modulating tau pathology: identifying a novel inhibitor for the Suv39h1 protein using a multi-omic, attention-based neural network approach to treat Alzheimer’s disease. Four stages of computer-aided drug design, from virtual screening through toxicity prediction. F5446 was identified as the lead inhibitor, with a binding affinity of −8.3 kcal/mol against the MAPT–Suv39h1 complex.

Research poster, page one
Introduction, methodology, and procedure.
Research poster, page two
Mathematical modeling, statistical analysis, and conclusions.

Full poster (PDF)↗

Projects

Alzheimer’s CNN
87% detection accuracy across 5,000+ imaging cross sections.
LRRC30
Predicted five functional networks for the gene in diabetes mellitus across three pathways.
Blueprint College Consulting
Founded; advising for 20+ clients.

Other