AI Researcher - Efficient AI (Contractor)
Core
Research and implement methods to make modern LLMs, VLMs, and AI agents faster, smaller, and more deployable on constrained devices.
Role type
Contract AI Researcher (Efficient AI)
Builds
Efficient AI models and inference pipelines for LG's future products (AI PCs, edge devices, robotics, intelligent vehicles)
Domain
Artificial Intelligence / Machine Learning / Systems Optimization
Deliverable
production ML models
Required skills
Python, PyTorch, LLM/VLM/Multimodal model development, model compression (PTQ, QAT, pruning), inference optimization, experimental pipeline design
Preferred skills
Publications in top ML venues, experience with inference frameworks (llama.cpp, vLLM, TensorRT-LLM), low-level kernel implementation, emerging architectures (MoE, SSMs)
Technologies
PyTorch, llama.cpp, GGUF, vLLM, SGLang, TensorRT-LLM, MoE, SSMs
Responsibilities
Research and prototype AI methods for model efficiency and inference performance; Optimize LLMs/VLMs across post-training and inference workflows; Propose and evaluate novel compression methods; Develop gradient-free methods for model merging and optimization; Implement emerging efficient architectures; Prototype inference-time optimization methods; Build experimental pipelines and evaluate on benchmarks; Contribute to publications and IP submissions
Seniority
Individual Contributor, Research-focused