À propos
Our client is a cutting-edge AI startup in the Bay Area developing highly efficient foundational models for real-world deployment across devices. Rapidly growing, highly technical team focused on building top-tier large language model (LLM) architectures with real-world impact. As a Member of Technical Staff, you’ll drive innovation on large-scale model training, infrastructure, and optimization. You’ll collaborate closely with a small team of seasoned researchers and engineers, advancing state-of-the-art LLMs for efficient deployment at scale. Responsibilities
Design, implement, and optimize large-scale pretraining and post-training pipelines for language models Tackle challenges in model parallelism, distributed training, and low-level hardware/software co-design Monitor, maintain, and troubleshoot massive training and inference workloads end-to-end Collaborate on advancing core model architectures, inference optimizations, and custom hardware design Contribute to open-source community initiatives and research publications Analyze and streamline data pipelines, instruction data curation, and evaluation methods Apply advanced optimization theory to improve model performance Qualifications
Degree in Computer Science, Electrical Engineering, or related technical field (or equivalent practical experience) Hands-on experience in machine learning research centered on LLMs, efficient AI systems, or large-scale model training Strong proficiency in Python and deep learning frameworks (e.g., PyTorch, TensorFlow) Expertise with distributed training, parallelization strategies, and large-scale computational infrastructure Understanding of low-level GPU optimizations, CUDA, or similar technologies Preferred Skills
Previous work at leading research labs or high-impact contributions to community AI projects Experience with custom hardware, FPGA/ASIC design, or maximizing training throughput Familiarity with open-source inference engines (e.g., llama.cpp, vllm, triton) Academic publications in optimization, LLM training, or AI infrastructure
#J-18808-Ljbffr
Compétences linguistiques
- English
Avis aux utilisateurs
Cette offre provient d’une plateforme partenaire de TieTalent. Cliquez sur « Postuler maintenant » pour soumettre votre candidature directement sur leur site.