Job description

This is a remote position.

Technical Approach - We have a working AI pipeline using text prompt optimization, text-to-image, and image-to-video using a diffusion model architecture
- We are using prompt optimization techniques and testing multiple open-source text-to-video models (OpenSora, Mora, IV2 Gen XL, etc) - We need to move into the fine-tuning approach using vector databases to isolate specific issues in videos (e.g., facial consistency, lighting conditions, camera angles, shutter speed, etc) Skill Requirements
- We need an experienced individual who approaches product development like a scientists
- This individual breaks down problems into small, isolated components, develops hypotheses to test and address these components, runs the experiments (in code), analyzes the outputs, and adjusts the next round of experiments accordingly.
- This person will develop an analysis framework, write code, and improve the underlying model.
- This person does not need to manage infrastructure or the web application but should focus on improving the output of the text-to-video model so that it can be used in an enterprise application.
- Strong understanding of computer vision algorithms in the context of video generation
- Experience working with diffusion models, transformers (NeRFs are a plus)
- Track record in implementing state-of-the-art research papers with reproducible results
- Experience in RLHF is a plus Experience
- The ideal candidate will have experience in working with generative AI to isolate issues and improve the output either in images or video Key Responsibilities
1. Prototype and report on top models or innovative approaches by iterating rapidly to deliver tangible business outcomes
2. Keep track of top research directions/papers in the field of video generative models to steam our innovation roadmap
3. Design and develop ML training pipelines for large-scale datasets
4. Collaborate with CTO and product development team in developing new feature

Salary:

2400000

Required profile