Visual Generative ML Engineer: Post-Training & Evaluation

Nunchux AI, Inc.; San Francisco (CA); Hybrid; USD 180,000 - 250,000; Full time

Benefits offered by this job

Health insurance

401(k) plan

Equity

Job summary

Nunchux AI, Inc. in San Francisco, CA is seeking a Machine Learning Engineer to build post-training pipelines that improve efficiency and quality of multimodal visual generation models.

You will develop post-training recipes, data pipelines, evaluation systems, and benchmarks, and bring cutting-edge research into production using Python and PyTorch. This is a hybrid role with four days in the office and one day remote.

Qualifications

• Hands-on experience with image or video generation models, or multimodal visual systems.
• Depth in post-training methods such as distillation, LoRA, or evaluation of visual generative models.
• Strong Python and PyTorch skills, with experience building post-training or evaluation code.
• Experience running post-training workloads across multiple GPUs with DeepSpeed or FSDP.
• Ability to design clean experiments, interpret results, and make data-driven recommendations.

Responsibilities

• Develop post-training recipes for image and video generation models.
• Build data pipelines to curate and version training and benchmark data.
• Create benchmarks and automated judges; run human preference studies.
• Benchmark and release models by measuring against baselines and identifying quality regressions.
• Integrate research methods into production pipelines and keep up with evolving post-training work.

Skills

Visual generative-model experience

Post-training or evaluation depth

ML engineering strength

Training systems

Experimental judgment

Tools

PyTorch

DeepSpeed

FSDP

Job description

Nunchux AI, Inc. in San Francisco, CA is seeking a Machine Learning Engineer to build post-training pipelines that improve efficiency and quality of multimodal visual generation models.

You will develop post-training recipes, data pipelines, evaluation systems, and benchmarks, and bring cutting-edge research into production using Python and PyTorch. This is a hybrid role with four days in the office and one day remote.