Experience

NODAR

• United States
• Software Development
• 1 - 100 Employee

• Principal Software Engineer - Computer Vision

  • Jun 2022 - Present



GE Baker Hughes

• Greater Boston

• Senior Staff Computer Vision Engineer

  • Jan 2020 - Jun 2022

  * built and managed our cloud-based autoscaling ML (machine learning) infrastructure in production with hundreds of CPU and GPU nodes in Kubernetes clusters; led the architecture, design, implementation and orchestration of our scalable ML infrastructure (AWS, GCP, and on-prem DGX clusters). * built a closed-loop ML annotation/training/validation/inference pipeline, which supported object classification, object detection, semantics segmentation, etc. It's cloud based and operated in highly-scalable mode. * built a large commercial-grade object detection and segmentation ML pipeline, which includes video stabilization, object detection and segmentation, KPI validation, etc. Again, it's cloud-based with full autoscaling capability, and the pipeline auto-scaled hundreds' of GPU and CPU cloud nodes on-demand. Show less



GE Ventures

• United States
• Venture Capital & Private Equity
• 1 - 100 Employee

• Staff Analytics Engineer (Computer Vision)

  • 2017 - 2020

  * Collaborated with NVIDIA, led and built first version of GE AI Workbench, a browser-based platform which allows customers to drag-and-drop to customize data, models and pipelines for no-code machine/deep learning training and inference purpose. * Built and managed our AWS cloud-based large-scale Structure from Motion (SfM) 3D modeling pipeline, which run on a daily basis and consumed TB of data per day. The pipeline has generated thousands of 3D reality models with drone/robotics images from our clients in production environment. * Developed a variety of computer vision and deep learning products, from object classification, component detection, change detection, video stabilization etc. * Led and built a full MVP product to classify flare images on the edge device (NVIDIA Jetson Xavier). The pipeline includes real-time image acquisition, concurrent and parallel processing design, deep learning based object classification. Show less



GE Research

• United States
• Research
• 700 & Above Employee

• Computer Vision Scientist

  • 2016 - 2017

  * Built and managed GE Research Computer Vision infrastructure from bare metal servers to Nvidia DGX clusters for training and inference. * Docker-containerized decade's GE Computer Vision codebase. By leveraging those created CV codebase Docker images, the new approach fundamentally changed and streamlined our development and production deployment pipeline by reducing time of new build/deployment from 2 weeks to less than one hour. * Delivered the second version of GE Aviation inspection pipeline, which includes the development of front end inspection web app, integration of GE CV LAB deep learning semantic segmentation models, and final build and integration with GE Predix platform. Show less



Independent Cloud Computing Consulting

• Albany, New York Area

• Full Stack Developer

  • 2014 - 2016



Nova Health Informatics Inc.

• Nashville

• Co-founder

  • 2010 - 2013



Independent Web and Computer Vision Consulting

• Nashville

• Full Stack Developer

  • 2004 - 2009



Nanyang Technological University Singapore

• Singapore
• Higher Education
• 700 & Above Employee

• Computer Vision Research Engineer

  • 2000 - 2003