Back to search

AI Engineer

Help train, fine-tune, and serve the generative models behind realtime, interactive AI avatars that can see, listen, and respond with natural speech and expressive motion.

  • Palo Alto, California, United States
  • Full-time
  • On-site
  • Posted Nov 17, 2026
INDUSTRYSoftware & InternetSaaS & Enterprise SoftwareAI & Machine Learning

Responsibilities

  • Train and fine-tune generative models for talking-head synthesis, audio-driven facial animation, and full-body avatar video
  • Build data pipelines for large-scale video, audio, and multimodal datasets
  • Optimize models for realtime serving, targeting sub-200ms end-to-end latency, using techniques such as distillation, quantization, and custom kernels
  • Build and maintain training, evaluation, and deployment infrastructure on GPU clusters
  • Define and improve metrics such as lip-sync accuracy, identity preservation, expression naturalness, latency, and throughput
  • Work with backend, product, and research teams to ship avatar features and iterate on user feedback

Requirements

  • Experience in a Physical AI

Nice to have: Talking-head systems, Lip-sync systems, 3D Gaussian Splatting, NeRF, Face reenactment

Skills

PyTorchDiffusion modelsVideo generationComputer visionMultimodal AISpeech synthesisNeural renderingTensorRTONNXTritonCUDAPythonMulti-GPU distributed trainingPyTorchTensorRTONNXTritonCUDA

Benefits

  • Competitive salary
  • Equity
  • Benefits
  • GPU compute access

More jobs nearby

Full description

About LiveX AI LiveX AI is a fast-growing startup based in the San Francisco Bay Area, building the next generation of physical AI agents—intelligent, embodied systems that interact with people in real-world environments. We combine cutting-edge language models, multimodal AI, and custom hardware to power transformative AI experiences in airports, stadiums, conferences, and more. We’re trusted by top enterprises and city partners, and backed by renowned investors. ABOUT THE ROLE We're building realtime, interactive AI avatars that can see, listen, and respond with natural speech and expressive motion. As an AI Engineer, you'll help train, fine-tune, and serve the generative models behind them—from data curation and training diffusion and video models to low-latency inference for live streaming. You'll work closely with our backend, product, and research teams to ship avatar experiences used in the real world. WHAT YOU'LL DO - Generative avatar models: Train and fine-tune generative models for talking-head synthesis, audio-driven facial animation, and full-body avatar video. - Data pipelines: Build data pipelines for large-scale video, audio, and multimodal datasets. - Realtime optimization: Optimize models for realtime serving, targeting sub-200ms end-to-end latency, using techniques such as distillation, quantization, and custom kernels. - ML infrastructure: Build and maintain training, evaluation, and deployment infrastructure on GPU clusters. - Quality metrics: Define and improve metrics such as lip-sync accuracy, identity preservation, expression naturalness, latency, and throughput. - Ship with the team: Work with backend, product, and research teams to ship avatar features and iterate on user feedback, while keeping up with the latest in generative video, multimodal AI, and neural rendering. WHO YOU ARE Must-haves - 3+ years of hands-on ML engineering experience, or strong research experience with production-quality code. - Deep PyTorch expertise and the ability to reproduce and extend research papers. - Background in diffusion models, video generation, computer vision, multimodal AI, speech synthesis, or neural rendering. - Experience optimizing realtime inference (e.g., TensorRT, ONNX, Triton, CUDA, mixed precision). - Strong Python skills and experience with multi-GPU or multi-node distributed training. - Comfort with ambiguity and fast-changing environments, with a strong sense of ownership. Bonus points - Experience with talking-head, lip-sync, or digital human systems, or related areas such as 3D Gaussian Splatting, NeRF, or face reenactment. - Publications at top venues (e.g., NeurIPS, CVPR, ICCV, SIGGRAPH, ICLR, ICML) or notable open-source contributions. WHY JOIN LIVEX AI? - World-Stage Experiences: Your work goes live in front of millions—at the Super Bowl, major airports, and NVIDIA GTC—not just inside an office. - Full-Real Scope: AI, hardware, and real customers—you'll build a skill set most people don't reach until years into their careers. - Silicon Valley, Front Row: Work at the center of the physical AI movement, with the connections and momentum that come with it. - Mentorship That Matters: Work directly with founders and senior leaders who want to see you grow—not hand you busywork. - Real Ownership & Growth: Top performers take on more scope fast—leading projects, owning customers, and shaping the product. - Compute to Build With: Access to GPU compute for training and experimentation, plus competitive salary, equity, and benefits. Ready to launch the next phase of your career and build the future of physical AI? Apply to join LiveX AI today.

AI Engineer at LiveX AI | Referso Jobs