Staff AI/ML Engineer – Large-Scale Systems
Hybrid (Pasadena or San Francisco)Full-timePosted 3mo ago
About the role
Staff-level AI/ML engineer to lead large-scale model training efforts, combining hands-on ownership of training runs with setting technical direction, mentoring engineers, and improving model quality and system performance across the organization.
Responsibilities
- Leading model development across pretraining, fine-tuning, and post-training stages
- Designing and improving data pipelines, including curation, filtering, deduplication, and dataset composition
- Improving training efficiency, scalability, and reliability across large distributed systems
- Optimizing model performance with respect to convergence, throughput, memory usage, and stability
- Translating cutting-edge research into robust, production-ready systems
- Providing technical leadership through mentoring, design reviews, and cross-functional collaboration
Qualifications
- 8–10+ years of experience in machine learning or AI or strong publication record
- Strong Python programming skills with production-quality code
- Hands-on experience training large-scale models (multi-billion parameters)
- Solid understanding of optimization, distributed training, and training dynamics
- Experience with modern model training workflows (e.g., pretraining, fine-tuning, reinforcement learning approaches)
- Proven ability to mentor and lead other AI/ML engineers
Skills
Required:Large-scale model trainingPythonOptimizationData pipeline designMentoring/leadership
Last updated Oct 22, 2026. Interested? Apply on the company site.
About PrismML
Concentrating intelligence
PrismML — Concentrating intelligence
Bonsai 27B Bonsai 8B Bonsai 4B Bonsai 1.7B Bonsai Image
More open roles at PrismML
- Senior AI/ML Engineer – Forward Deployment2mo agoHybrid (Pasadena; San Francisco) · PrismML Careers / Engineering
- Senior AI/ML Engineer – Kernel Optimization3mo agoHybrid (Pasadena;San Francisco), Remote · Engineering
- Senior AI/ML Engineer – Post-Training Platform3mo agoHybrid (Pasadena or San Francisco)