Realtime AI Avatar Engineer — Multimodal & Low-Latency

Embedding VC; Palo Alto (CA); On-site; USD 140,000 - 230,000; Full time

Job summary

LiveX AI in Palo Alto is seeking an experienced AI Engineer (Machine Learning Engineer) to train, fine-tune, and deploy generative models for talking-head avatars, including diffusion and video generation, with emphasis on low-latency inference.

You will build data pipelines, optimize multi-GPU training, and collaborate with backend, product, and research teams to ship avatar features while staying current with advances in multimodal AI and neural rendering.

Qualifications

• 3+ years of hands-on ML engineering or research with production-quality code.
• Deep expertise in PyTorch and modern DL workflows.
• Experience with diffusion models, video generation, or neural rendering.
• Strong Python engineering and distributed training on multi-GPU / multi-node setups.

Responsibilities

• Train and fine-tune generative models for talking-head synthesis, audio-driven facial animation, and full-body avatar video.
• Design data pipelines for large-scale video, audio, and multimodal datasets: collection, cleaning, labeling, alignment, augmentation.
• Optimize models for realtime serving with sub-200ms end-to-end latency: distillation, quantization, pruning, caching, streaming inference, custom CUDA/Triton kernels where needed.
• Build and maintain training, evaluation, and deployment infrastructure on GPU clusters (PyTorch, DeepSpeed, FSDP, Ray, Kubernetes).
• Define and track quality metrics (lip-sync accuracy, identity preservation, expression naturalness, latency, throughput) and drive measurable gains.
• Work with backend, product, and research teams to ship avatar features and iterate on user feedback.
• Keep up with research in generative video, multimodal AI, and neural rendering; bring useful techniques into the stack.

Tools

PyTorch

TensorRT

CUDA

ONNX

Kubernetes

Job description

LiveX AI in Palo Alto is seeking an experienced AI Engineer (Machine Learning Engineer) to train, fine-tune, and deploy generative models for talking-head avatars, including diffusion and video generation, with emphasis on low-latency inference.

You will build data pipelines, optimize multi-GPU training, and collaborate with backend, product, and research teams to ship avatar features while staying current with advances in multimodal AI and neural rendering.