Machine Learning Engineer – Video Diffusion Models

Lit8 develops generative AI systems for real-time video generation and enhancement. In this role, you will develop, train, and optimize advanced video generative models, focusing on video diffusion, motion realism, temporal consistency, controllability, and inference performance.

We are looking for someone with substantial hands‑on experience training and improving diffusion models, including video generation systems, with a strong understanding of temporal modeling and large‑scale generative architectures.

Minimum Qualifications

• At least 4 years of hands‑on experience in machine learning or deep learning.
• At least 3 years of hands‑on experience with diffusion models, including training, fine‑tuning, evaluating, or optimizing them.
• At least 2 years of hands‑on experience with video diffusion models.
• Strong experience with video generation, including image‑to‑video, text‑to‑video, video‑to‑video, or related generative tasks.
• Understanding of temporal consistency, motion realism, identity preservation, flicker, and long‑range video coherence.
• Experience with diffusion training, sampling, conditioning, guidance, and latent generative architectures.
• Strong Python programming skills and proficiency with PyTorch.
• Experience evaluating video generation quality, including temporal stability, motion quality, perceptual metrics, and artifacts.
• Strong analytical and problem‑solving skills.

Preferred Qualifications

• Experience with autoregressive (AR) video models, video transformers, or hybrid autoregressive/diffusion architectures.
• Experience with GAN‑based image or video generation models.
• Experience with diffusion transformers (DiTs), flow matching, rectified flow, or other advanced video generative architectures.
• Experience training large‑scale video models using distributed training, mixed precision, and extensive video datasets.
• Experience optimizing video models through distillation, quantization, reduced‑step sampling, or hardware‑aware inference.
• Contributions to open‑source video generation projects or relevant research publications.

Key Responsibilities

• Develop, train, and optimize video diffusion models for high‑quality video generation.
• Improve motion realism, temporal coherence, visual quality, and controllability.
• Develop image‑to‑video, text‑to‑video, and video‑to‑video generation capabilities.
• Improve training pipelines, sampling methods, and evaluation workflows.
• Collaborate with research and engineering teams to integrate video generative models into production applications.

What We Offer

• The opportunity to work on advanced video generative AI systems with real product impact.
• A research‑driven environment focused on technical excellence and ownership.

If you are passionate about video diffusion, temporal generative modeling, and building next‑generation video generation systems, we'd love to hear from you.