Implement AI/ML into SAAS app using AWS Rekognition

Use AI and Machine learning (AWS…Rekognition) to automatically tag image and video files when they are uploaded into the application. ML should allow for learning how how an account is tagging their files and then use that methodology automatically going forward. Each customer account will have separate information/methodology -so the machine needs to learn at the account level not generally across the application. Users will still be able to manually tag files and filter by manual tags….but also use AI and filter by AI generated tags. Current Tech Stack • AWS (S3, EC2, RDS) • Node.js (Express) • React (next) Back end Details • AI Turned on for a customer or turned off • Unique machine learning for each customer (that has AI turned on) • Can learn from tags, campaigns, collections, custom fields (these are current metadata “tags” used in the application • Apply learnings to new uploads • Run process on existing assets that were uploaded prior to AI tagging • Option for user to accept or reject AI tags • Manual training of model (rules)….filename structure, custom field to tag correlations. Front End (optional….can use other dev to do front end if needed) Allow admin to turn on or off for an account Front end filter showing new AI tags Accept/Reject proposed tags (pop up modal with ai generated tags) Manual model training Manual Rules (If file name contains “X”, then “Y” tag) Select assets and run AI tag process on files already in system (load mgt)