Akshay Ashok Nair

Trained a CellCNN-style CNN on 2,000+ patient samples (up to 8,192 cells/patient) to predict ER status from flow cytometry data, achieving AUROC 0.73 on a held-out test set Built a dual-backend architecture with a shared data pipeline feeding both a Keras CNN and a CytoDx R model Ensured reproducibility via stratified train/val/test splits, z-score normalization, and parameter-encoded disk caching

Research Assistant - Foy Lab - University of Washington - Seattle, WA

(2026-01)

• Trained a CellCNN-style CNN on 2,000+ patient samples (up to 8,192 cells/patient) to predict ER status from flow cytometry data, achieving AUROC 0.73 on a held-out test set
• Built a dual-backend architecture with a shared data pipeline feeding both a Keras CNN and a CytoDx R model
• Ensured reproducibility via stratified train/val/test splits, z-score normalization, and parameter-encoded disk caching

Project Lead - The Boring Club (Face Recognition System, LiDAR Maze Navigation) - University of Washington - Seattle, WA

(2025-09)

• Directed 15 student engineers across two hardware/ML projects, defining system architecture and weekly milestones
• Made key design trade-offs for edge deployment on constrained hardware, achieving 95% accuracy on both systems

ML Researcher - Burkitt's Lymphoma Fund for Africa - Seattle, WA

(2021-06 - 2024-06)

• Developed and fine-tuned CNNs for Burkitt's Lymphoma detection on histopathology slides, achieving 90%+ accuracy
• Performed image preprocessing, data analysis, and model evaluation using Python and PyTorch
• Collaborated with UW, Fred Hutch, and Phenopath Labs on system design, validation - led to first-author publication

Research Assistant - Sensors, Energy, and Automation Lab - University of Washington - Seattle, WA

(2024-04 - 2025-09)

• Built JavaScript and Python/Selenium automation tools to streamline member tracking and research data workflows
• Collaborated with graduate researchers to prepare a research manuscript for submission to the CDC

Software Development Intern - Liberated Logic - Bothell, WA

(2023-06 - 2023-09)

• Designed a user-centered prototype website and built interactive Figma wireframes to improve early-stage usability
• Strengthened front-end development skills through rapid prototyping with React and component-based design

Bachelor of Science - Electrical and Computer Engineering - University of Washington (2027-06)