About Unfaze.ai

Create stunning product images and short-form promotional videos 10x faster.

Tech description:

In addition to utilizing text-to-image foundation models, we train new kinds of controlnets, LoRAs, upscalers, and other proprietary machine learning models. The machine learning framework we use is Pytorch. Our tech stack is RedwoodJS (React, Prisma, Postgres, GraphQL, Storybook,), TailwindCSS, Supabase, and FastAPI. Our primary technical objective is to tame the beast that is the text-to-image diffusion model by establishing fine-grained control over the image generation process, immaculate photorealism, and superhuman prompt-understanding. Achieving these goals involves solving novel machine learning challenges that push beyond the existing research frontier. For those initiated in text-to-image diffusion models, challenges like preserving text, fine-grained details, improving image photorealism, and achieving accurate multi-subject compositions in raw generated outputs will be familiar. We tackle these challenges every day, and want to work with ML Researches who are similarly excited about solving them! Commercial photography is undergoing massive change. Using Latent Studio, any business can conduct a professional photoshoot workflow from their computer, produce images that match the quality of those used by the biggest brands in the world, and scale their content creation process at virtually zero marginal cost.

Job description:

Latent Studio makes the creation of professional quality product photos 10x cheaper & faster for e-commerce businesses and DTC brands through the use of text-to-image diffusion models. We’ve raised $3.4mm in funding from investors like Hack VC, Soma Capital, and Pioneer Fund, and are currently a team of 7. Our primary technical objective is to tame the beast that is the text-to-image diffusion model by establishing fine-grained control over the image generation process, immaculate photorealism, and superhuman prompt-understanding. Achieving these goals involves solving novel machine learning challenges that push beyond the existing research frontier. For those initiated in text-to-image diffusion models, challenges like preserving text, fine-grained details, improving image photorealism, and achieving accurate multi-subject compositions in raw generated outputs will be familiar. We tackle these challenges every day, and want to work with ML Researches who are similarly excited about solving them! Commercial photography is undergoing massive change. Using Latent Studio, any business can conduct a professional photoshoot workflow from their computer, produce images that match the quality of those used by the biggest brands in the world, and scale their content creation process at virtually zero marginal cost.

Skills:

Python, Machine Learning, Computer Vision

Unfaze.ai remote jobs

ML Researcher