Engineer, Senior Staff
Core
Train Large Language Models (LLMs) from scratch and fine-tune existing LLMs on various datasets using state-of-the-art techniques for Qualcomm's Cloud AI team.
Role type
Senior Staff Language Model Engineer
Builds
Machine Learning solutions for data center, edge, infrastructure, and automotive markets
Domain
Artificial Intelligence / Natural Language Processing
Deliverable
production ML models
Required skills
PyTorch, Python, Large Language Models (LLMs), Natural Language Processing (NLP), Model Architecture Optimization, Reinforcement Learning from Human Feedback (RLHF), ZeRO, Speculative Sampling, Data Cleaning, Model Evaluation, ONNX optimization
Preferred skills
Quantization (AWQ), Quantization Aware Training (QAT), Neural Architecture Search (NAS), CUDA, CUDNN, Triton-lang, Data version control
Technologies
PyTorch, ONNX, ZeRO, RLHF, Triton-lang, CUDA, CUDNN
Responsibilities
Optimize LLM and GenAI model architectures for NPUs by reimplementing basic building blocks; Fine-tune pre-trained models on specific tasks or datasets; Handle large datasets ensuring data quality and integrity; Evaluate model performance using appropriate metrics; Stay up to date with latest research in NLP and LLMs; Develop new optimization techniques to minimize ONNX memory footprint and export time
Seniority
Senior Staff, 4-10 years experience (preferred 9-12 years)
