Lead data scientist developing computer vision applications for cloud and edge.
• Engineered SOTA auto-labeling and ModelOps tools, accelerating model development flow and reducing personnel costs.
• Enhanced model efficacy through advanced techniques like Genetic Algorithms, Hyperparameter Optimization, and adaptive model architecture adjustments, including graded learning rates and selective head freezing.
• Architected, trained, and validated custom vision models from scratch.
• Monitored and maintained production models against degradation and drift.
• Built critical monitoring dashboards, saving thousands a month in cloud costs and exposing breaking software bugs.
• Constructed and deployed a Retrieval Augmented Generation tool making company documentation more discoverable and reducing personnel hours.
• Represented data team in customer on-sites and conclusive presentations leading to pilots, jumpstarts, and product rollouts.

Full stack dev for secure internal website and software dev for ML lab. Responsibilities included building webpages and data GUIs, SQL database migration and management, and constructing and executing unit tests on key landing pages. See public website here.

Developed a robust multimodal deep-fusion autonomous vehicle model for object detection in adverse weather conditions. See project details here.