Machine Learning Engineer – Model Distillation for Real-Time 2D Image Generative A

Lit8 develops generative AI systems for

real-time 2D image generation and enhancement . In this role, you will focus on distilling, optimizing, and deploying high-performance image generative AI models, with an emphasis on speed, quality, controllability, and production-ready performance.

You will work closely with research, engineering, and product teams to make advanced image generation models faster, lighter, and suitable for real-world applications.

Minimum Qualifications

At least

2 years

of hands‑on experience distilling image generation models, preferably image-to-image models.

Strong experience with

2D image generative AI , including diffusion models, transformer‑based image models, GANs, or other generative architectures.

Practical experience with model distillation techniques such as teacher‑student training, progressive distillation, consistency distillation, adversarial distillation, feature‑level distillation, score distillation, or latent‑space distillation.

Experience working with image-to-image generation tasks such as inpainting, outpainting, super‑resolution, denoising, image editing, style transfer, enhancement, or controllable generation.

Hands‑on experience training, fine‑tuning, evaluating, and optimizing image generation models.

Experience improving inference latency, memory efficiency, throughput, and model quality.

Strong programming skills in Python.

Hands‑on experience with modern ML frameworks, especially PyTorch.

Solid understanding of model compression, mixed precision, quantization‑aware optimization, pruning, or related efficiency techniques.

Strong problem‑solving, analytical, and communication skills.

Ability to work effectively in a fast‑paced, research‑driven, multidisciplinary technical environment.

Preferred Qualifications

Experience deploying optimized generative AI models into production applications, device‑specific pipelines, or consumer‑facing products.

Familiarity with inference and deployment frameworks such as ONNX, TensorRT, OpenVINO, Core ML, DirectML, ROCm, Vulkan, or similar technologies.

Experience benchmarking generative AI systems, including latency, throughput, memory usage, image quality, visual consistency, and stability.

Experience with multimodal or foundation models for image generation, editing, enhancement, or controllable visual generation.

Knowledge of GPU performance optimization, custom kernels, operator fusion, graph optimization, or hardware‑aware model tuning.

Contributions to open‑source ML, computer vision, image generation, or model optimization projects are a plus.

Relevant publications or research experience in generative AI, computer vision, model compression, or efficient inference are a plus.

Key Responsibilities

Develop and apply

model distillation techniques

to accelerate 2D image generative AI models.

Work on image‑to‑image and related generative AI workflows, including editing, enhancement, denoising, inpainting, and super‑resolution.

Improve model efficiency while preserving image quality, controllability, visual consistency, and robustness.

Train, fine‑tune, and evaluate distilled models across different image generation tasks.

Prototype and benchmark distillation strategies across different architectures and deployment targets.

Optimize inference performance through distillation, compression, quantization, mixed precision, pruning, graph optimization, and memory‑aware tuning.

Build evaluation workflows to measure model quality, latency, memory usage, throughput, and reliability.

Collaborate with research, engineering, and product teams to integrate optimized models into production applications.

Stay current with advances in image generative AI, model distillation, efficient diffusion models, and real‑time inference.

What We Offer

The opportunity to work on advanced

real‑time 2D image generative AI

systems.

A fast‑moving, research‑driven environment with real product impact.

The chance to make state‑of‑the‑art image generation models faster, lighter, and production‑ready.

A culture that values technical excellence, ownership, creativity, and performance engineering.

If you are passionate about image generative AI, model distillation, and building efficient production‑grade AI systems, we’d love to hear from you.

#J-18808-Ljbffr