Alexander Ponamarev

Senior Machine Learning Engineer (Deep Learning) @Artisight

Austin, TX, US

Delivered deep learning (Computer Vision and LLM) systems to enhance patient outcomes. Own full AI product life-cycle, including requirements gathering, managing data labeling process, setting up and maintaining CICD pipelines.• Designed, implemented, and delivered LLM based services - Patient Length-of-Stay and Healthcare Resource Utilization. Services were deployed at a large hospital network supporting over 600 patients on a daily basis. • Led team of 4 engineers to set up data collection, train LLMs (via parameter efficient fine tuning), implement and deploy model inference service.• Developed and deployed Surgical-Action Recognition service (computer vision) in Northwestern University hospital to provide support for all laparoscopic surgeries.• Created a model training process (PyTorch) that used parameter efficient fine tuning (LoRA) to train large vision models (DINOv2) and distill the resulting model into a light production model (EfficientNet Small). This training process increased model accuracy (+7%) while maintaining high inference speed (30+ FPS).• Developed self-supervised learning (SSL) pipeline that reduced manual labeling needs by 80% and increased product accuracy (+9%).• Deployed object detection (YOLO) and action recognition (EfficientNet backbone + LSTM) models into production responsible for +20% of AI revenue of the company.

Udacity

Computer Vision Nanodegree, Computer Science

Udacity

AI Nanodegree, Artificial Intelligence

Alliance Manchester Business School

Master of Business Administration (M.B.A.), Business Administration and Management, General