// resume

Engineering Resume

Computer vision, machine learning, and AI from early concept to production: marine perception, autonomy, cloud-to-edge pipelines, and IP generation.

Download as PDF

01 Summary

Computer vision engineer who researches, develops, and deploys computer vision, machine learning, and AI solutions end to end, from early concept to production. At Brunswick's Boating Intelligence Design Lab, I lead CVML R&D for the recreational marine industry, taking work from proof of concept to a model deployed in a Level 3 autonomous docking system. I have experience building large-scale datasets, training and evaluating deep learning models, architecting cloud-to-edge ML pipelines, and applying multimodal AI to operational problems. Inventor on more than fifteen U.S. patent filings, and driven to lower the barrier to entry into boating by building ADAS solutions that make boating easier and safer for boaters of every experience level.

02 Work Experience

Jul 2021to Present
Technical Associate, Computer Vision
Brunswick Corporation, Boating Intelligence Design Lab
  • Lead Brunswick's computer vision R&D, developing solutions such as vessel blind-spot monitoring and obstacle avoidance; build proofs of concept to validate new product ideas, present them to senior leadership, and help shape the organization's product roadmap.
  • Filed over fifteen U.S. patent applications in computer vision and AI, two of which have been granted, with the rest patent pending, and authored several dozen invention disclosures.
  • Architected and led development of the end-to-end production computer vision pipeline for Brunswick's Level 3 Autonomous Docking Solution, integrating cloud AI/ML services and hardware for data collection, labeling, storage, training, evaluation, and edge deployment.
  • Led one of Brunswick's operational-excellence AI initiatives, developing a multimodal AI solution for nautical chart data collection, extraction, formatting, and map generation; reduced processing time per data source from several days to a few hours with a fraction of the previous headcount.
  • Spearheaded the large-scale Marine Vision dataset development, collecting, processing, and labeling hundreds of thousands of images with external collaborators; built models across image classification, object detection, pose estimation, and segmentation.
  • Developed Brunswick's vision model evaluation methodology and explainability approach to assess production readiness, validating a model deployed in a Level 3 autonomous vessel auto-docking system.
  • Led an AI-driven CAN bus reverse-engineering project to test the security capabilities of Brunswick's proprietary CAN network.
  • Won the Brunswick Navico Division Hackathon with a concept for automated alignment of nautical chart data with marine radar imagery.
  • Transitioned R&D proofs of concept to production by partnering across divisions and business stakeholders to understand end-user pain points; presented solutions to senior leadership, including the CEO, for production approval.
  • Mentored and managed over a dozen research students, from freshmen to PhDs, from the University of Illinois Urbana-Champaign, and led collaborations with MIT's SeaGrant Lab and UMass Lowell.
Jan 2020to May 2021
Robotics Teaching and Research Assistant
Department of Electrical & Computer Engineering, UMass Lowell
  • Developed marine perception ontologies and COLREGS-based object detection models for the perception system of an autonomous prototype vessel, "Philos," a Boston Whaler Outrage 350; a joint project with MIT's SeaGrant Lab, funded by Lockheed Martin.
  • Built and integrated an object detection model for marine vessels and obstacles using ROS, Python, and the YOLO Darknet library.
  • Served as teaching assistant for the undergraduate robotics course, leading lab sessions and mentoring students on robotics curriculum and projects using ROS 1 and ROS 2.