Unconventional AI
Member of Technical Staff - AI Training Platform
Mountain View, CA or Greater Seattle Area +1 · Staff · Full-time · Posted Sep 2, 2026
- Location
- Mountain View, CA or Greater Seattle Area
Mountain View - Experience
- 5+ years of experience in AI/ML engineering.
- Education
- BS in Computer Science, Physics, Electrical Engineering, or Applied Math.
About the job
Build and scale the infrastructure, frameworks, and tools for training, evaluating, benchmarking, and optimizing AI models on novel hardware.
What you'll do
- Architect and maintain distributed AI/ML/RL training infrastructure, including data streaming, sharding, checkpointing, and recovery.
- Develop and optimize training frameworks, kernels, and benchmarking tools for model and hardware performance.
- Build hardware evaluation and model-optimization tools, collaborating with researchers and engineering teams.
What they ask for
- BS in Computer Science, Physics, Electrical Engineering, or Applied Math.
- 5+ years of experience in AI/ML engineering.
- Strong programming skills in Python or C++ and a track record maintaining production-grade training frameworks.
- Expertise in distributed model partitioning, communication primitives, parallelism strategies, and mapping AI architectures to system performance.
Nice to have
- MS/PhD or equivalent research/project experience in AI/ML or high-performance computing, with publications.
- Experience training or post-training large-scale LLMs and generative models; experience with listed technologies such as Kubernetes, CUDA, Triton, NCCL, vLLM, SGLang, VeRL, or TRL.
Tools
Python, C++, Megatron-LM, DeepSpeed, Ray, PyTorch Lightning, CUDA, Triton
AI focus
Develop infrastructure and tools for model training, evaluation, benchmarking, hardware mapping, and model optimization for performance and energy consumption.
Model-written summary · the original posting is authoritative