At Playground, we imagined and built next-generation tools for creating and editing graphics using AI.
Some highlights of my work:
• Built image rendering and compositing pipelines for an infinite-canvas graphics editor
• Built a system for managing concurrent asynchronous image generations
• Built, tuned, and scaled a system for classifying images into categorized "Rising" feeds using CLIP
• Implemented notifications with smart grouping for interactions on user-generated content (likes, follows)
• Productionized and scaled from UI to inference: ControlNet, outpainting, Dreambooth, etc.
• Fixed artifact-causing issues in sampling algorithms in HuggingFace's open-source diffusers library
• Built a system for showing an animation of intermediate steps during image generation
• Tamed Google Cloud services to be cost efficient at scale
• Interviewed candidates, helped onboard+mentor new hires, helped with planning and prioritization, etc.
