Computer Vision Consultant

Werkr Tractors, December 2024 - Current

• Leading the development of a proof-of-concept computer vision system for autonomous tractor navigation, designing solutions for challenging agricultural environments and varied terrain conditions.
• Implementing novel integration between custom vision models and Robot Operating System (ROS) framework, enabling robust autonomous navigation capabilities for agricultural applications.

Co-chair

Conference on Vision and Intelligent Systems, October 2024 - January 2025

• Led the organization of a 100+ attendee conference, working with a team of 4 committee members and coordinating the review of 50+ technical papers.
• Established partnerships with 5+ major tech companies, securing $5K in sponsorships to support conference activities and student participation.

Research Project Manager

Miovision, August 2022 - February 2023

• Managed and coordinated the NSERC/OCI-funded project “Traffic Safety Margin Inference via 3D Spatial Modeling” in collaboration with the University of Waterloo Vision and Image Processing Lab.
• Oversaw project planning, research execution, and stakeholder communication, ensuring alignment with research milestones. Facilitated data collection, model development, and validation for 3D spatial traffic safety analysis while liaising with academic and industry partners to drive innovation and application.

Computer Vision Developer

Miovision, March 2020 - February 2023

• Improved production model accuracy by 12% (mAP) through implementation of an advanced traffic object detection (SSD) system while maintaining backward compatibility.
• Led the successful migration of the entire training pipeline to the latest TensorFlow version, reducing training time by 8% and improving model deployment efficiency.
• Developed and deployed robust validation procedures that reduced false positive rates by 5% in challenging weather conditions

Head Researcher

AICAT, January 2019 - December 2019

• Developed and implemented real time polyphonic piano note transcription model with PyTorch, Librosa and SoundFile based on RCNN.
• Helped raise 80% of the company’s funding (~$200k USD) by winning the NAVER AI-Starthon.
• Spoke at international conferences with 6000 + attendees about AI’s role in international patents and art.

Senior Researcher & Consultant

SK Hynix, June 2018 - November 2018

• Designed and implemented a semiconductor grade prediction model using Keras.
• Built a custom Python generator function to solve data binary class imbalance problem.
• Utilized DeepExplain library to analyze data features with high grade correlation.
• Developed advanced statistical analysis skills by working with pre-processing and post analysis.

Home