Machine Learning Engineer – GAN-Based Image Generative AI

Lit8 develops generative AI systems for real-time image generation and enhancement. In this role, you will develop, train, and improve GAN-based image generation models, focusing on visual quality, training stability, controllability, and production-ready performance.

We are looking for someone with deep hands‑on GAN expertise — someone who has trained GANs, debugged adversarial training, and improved generator-discriminator systems in practice, rather than simply using pretrained models.

Minimum Qualifications

• At least 3 years of hands‑on experience with GANs, including training, fine‑tuning, debugging, and optimizing GAN‑based image generation models.
• Strong practical experience with generator-discriminator training, adversarial losses, training stability, and mode collapse mitigation.
• Experience building image generation, image‑to‑image, super‑resolution, enhancement, or related visual generation systems.
• Hands‑on experience designing training pipelines, improving loss functions, and evaluating generative models.
• Strong understanding of image quality evaluation, including FID, LPIPS, SSIM, perceptual quality, and visual artifacts.
• Strong Python programming skills and proficiency with PyTorch.
• Strong analytical and problem‑solving skills.

Preferred Qualifications

• Experience with large-scale GAN architectures, such as GigaGAN, or other high‑resolution image generation systems.
• Experience training GANs on large datasets using distributed training, mixed precision, and scalable data pipelines.
• Experience distilling diffusion models into GAN-based generators or combining adversarial training with diffusion distillation.
• Experience optimizing GANs for low‑latency or production inference.
• Knowledge of quantization, pruning, operator fusion, or hardware‑aware optimization.
• Contributions to open‑source GAN or image generation projects.

Key Responsibilities

• Develop, train, and optimize GAN-based image generation models.
• Improve image realism, visual quality, training stability, and controllability.
• Debug and refine adversarial training, loss functions, and data pipelines.
• Evaluate architectures and benchmark image quality and inference performance.
• Collaborate with research and engineering teams to integrate models into production applications.

What We Offer

• The opportunity to work on advanced GAN-based image generation systems with real product impact.
• A research‑driven environment focused on technical excellence and ownership.

If you are passionate about GANs and have substantial experience training high-quality generative models, we'd love to hear from you.