About the job
Core
Develop domain-specific sLLMs and VLMs, build and operate fine-tuning pipelines and serving systems, and optimize inference performance.
Role type
Senior IC machine-learning engineer (LLM fine-tuning & deployment)
Builds
Production-ready domain-specific AI models and inference serving infrastructure
Domain
Generative AI, Large Language Models, Multimodal AI
Deliverable
production ML models
Required skills
PyTorch, LLM fine-tuning (LoRA, SFT, DPO), multi-GPU training, model evaluation design, model quantization, inference optimization
Preferred skills
Reinforcement learning-based post-training, VLM/multimodal model architecture, Kubernetes-based learning/serving, NPU/constrained environment deployment, open-source contributions, global communication
Technologies
PyTorch, LoRA, SFT, DPO, Kubernetes, NPU
Responsibilities
Develop domain-specific sLLMs and VLMs from dataset construction to evaluation; Build and operate fine-tuning pipelines and serving systems; Perform model quantization, quantization, and inference performance optimization; Validate new open-source models and accelerators to establish application criteria
Seniority
Senior, hands-on IC