Aaron Silverberg

2 Projects

Flexible Vision
A SASS platform for TensorFlow annotations and iterative training

Automate quality assurance with image detection
  • Created several custom Node-Red modules for configuring onprem cameras and communicating with Cloud API.
  • Added API endpoints to Flask API for manipulating images using any OpenCV method
  • Wrote ExpressJS server as proxy service to up / download photos through Minio (AWS S3) and other storage services
  • Created Drop-n-Drop photo library manager for organizing training data from remote and local photos
  • Customized Tree component for infinite nesting of tag hierarchies within a photo.
  • Built client side analysis methods for easier review of tags across all photos, showing totals and average region sizes
  • Customized drawing component for creating, dragging, and resizing tags with 0 - 400% magnification
  • Structured dataset to provide tag categories, notes, and names, aside from types
  • Wrote client-side demo builder to double as a unit testing for reloading randomized training tags customizing parameters

Started: 2018-05-01

Launched/Lasted: 2023-12-31

Technologies: Node.js, Express.js, ReactJS, AWS S3, Docker, Minio, TensorFlow, Kubernetes, Google Cloud Platform, Node-Red, Python, Flask, OpenCV

Industries: Technology, Analytics, Robotics, Manufacturing

Team: E.A.Taylor: Front-End & API. Sam Silverberg: Back-End. Alec Goeckel: Back-End. Aaron Silverberg: Product Development.

Companies/Brands: Flexible Assembly Systems, Refined Motion Engineering

Refined Motion

Assembly line automation

HVFIR Project: Torque and Ohm Resistance Validator

  • Built Angular / Node-Red interface to tell direct operator to each fastener, in their proper order, & indicate when in-position to trigger the tool, and whether results were a pass or fail
  • Wrote strict data validation on torque and resistance values in order to allow operator to proceed or readjust fastener.
  • Wrote MySQL queries to get required settings for each step of the assembly process
  • Wrote MySQL queries to record results of each attempt to fasten

Assembly line job builder and progress monitoring

  • Built offline-first interface to configure assembly-line jobs and control tools & machinery over WebSockets and LocalStorage
  • Deployed Node-Red backend on Raspberry PI connected to station Windows computer
  • Built client-side performance tracker averaging the time to assemble one project in a job
  • Created custom node-red modules for smoother synchronization with Fanuc robots
  • Wrote Station Service request tool to allow operators to order more parts or tool servicing
  • This project expanded on the HVFIR project above by allowing management to configure and validate jobs manually without needing the MySQL database

Started: 2017-12-01

Launched/Lasted: 2018-04-01

Technologies: Node-Red, Node.js, JavaScript, HTML5, ReactJS, Express.js, Bootstrap, WebSockets

Industries: Technology, Robotics, Manufacturing

Team: E.A.Taylor: Front-End, Back-End. Aaron Silverberg: Vice President. Saiman Shetty: Project Management

Companies/Brands: Tesla Motors, Flexible Assembly Systems, Refined Motion Engineering

Other team