Jan 2023 – Jan 2025 · 2 yrs

• Developing of cross-platform safety-critical application features for robotic-assisted navigation systems in a product forward team.
• Mako SmartRobotics team assisting surgeons to optimize implant placement through 3D CT-based planning and precise haptic technology.

Jan 2020 – Jan 2023 · 3 yrs · Greater Seattle Area

• Software engineer designing, building and improving scalable backend services and systems as part of the Google Compute Engine, GCP team
• Lead contributor in reduced GCE hardware turnup deployment time, saving $X M per annum in revenue opportunity and hardware amortization costs.
• Responsible for end to end software development life cycle- driving design, implementation, integration testing, monitoring and production launch of reliability focused projects in the team.
• A champion for code health - lead contributor in analysis of Java end-to-end tests targeting engineering productivity and minimizing technical debt.
• Achieved 4x reliability, improved flakiness, and developer velocity via refactored large unit tests.
• Initiatives recognized by the Healthys Silver Award - Google wide global award to highlight code health improvement efforts. Additional recognition through two spot bonuses for impactful reduction in number of test timeouts.

Jan 2017 – Jan 2020 · 3 yrs · San Francisco Bay Area · On-site

• Research engineer in the Highly Automated Driving team. Developed efficient computer vision algorithms for camera and LiDAR data to create high resolution maps.  Development in C++ and Python and deploying at scale with AWS.
• Privacy Blurring : face and license plate detection
• GDPR compliance driven project to blur identifiable entities such as faces and license plates from images. Designed a supervised deep learning based object detector.
• End-to-end execution and project lead for input data selection, feature engineering, machine learning model architecture, training and model evaluation.
• Designed and leveraged YOLO and YOLOv3 (end to end convolutional neural networks) for detecting tiny faces in high resolution images.
• Epipolar error metric to measure camera pose (rotation and translation) errors
• Targetted a testing metric to quantify camera poses errors.
• Formulated image correspondence data from 3D survey ground control points to yield an epipolar error between camera poses.
• Improved development cycle of the team and set-up framework for regression testing.
• Led the Women's Initiative Network (WIN) ERG Berkeley chapter in 2019.