Efficient AI Researcher — On-Device & Multimodal LLMs

LG Electronics; Santa Clara (CA); Hybride; USD 116 000 - 125 000; Plein temps

Avantages offerts par ce poste

No-cost medical premiums

401(k) with matching

Paid time off

Résumé du poste

LG Electronics in Santa Clara, CA, is seeking a Contract AI Researcher - Efficient AI to join the Emerging Technology Lab. You will research and prototype methods to improve model efficiency, enabling faster inference on constrained devices and deployment in real-world workflows.

Ideal candidates have strong ML or AI systems experience, with Python and PyTorch proficiency, and a track record in LLM/ multimodal model work.

Qualifications

• MS/PhD in CS, CE, ML or related field.
• Experience in ML or AI systems and model optimization.
• Proficiency in Python and PyTorch.
• Experience with LLMs/SLMs/VLMs and generative AI systems.
• Ability to read papers, run experiments and report results clearly.
• Excellent written and verbal communication; adaptable in fast-paced env.

Responsabilités

• Research, prototype and implement AI methods to improve model efficiency.
• Optimize LLMs/SLMs/VLMs and agentic workloads for inference and deployment.
• Propose and evaluate compression methods for on-device LLM/VLMs.
• Address long-context inference and KV cache compression challenges.
• Develop gradient-free methods for model merging and efficiency optimization.
• Prototype inference-time optimization and kernel-level improvements.
• Build experimental pipelines and evaluate on language, vision and reasoning benchmarks.
• Contribute to publications, reports, open-source releases and IP.

Connaissances

Python

PyTorch

ML research

LLMs/Multimodal

Communication

Formation

MS/PhD in CS/ML/related field

Outils

llama.cpp

GGUF

vLLM

TensorRT-LLM

Description du poste

LG Electronics in Santa Clara, CA, is seeking a Contract AI Researcher - Efficient AI to join the Emerging Technology Lab. You will research and prototype methods to improve model efficiency, enabling faster inference on constrained devices and deployment in real-world workflows.

Ideal candidates have strong ML or AI systems experience, with Python and PyTorch proficiency, and a track record in LLM/ multimodal model work.