Distributed Training and Inference Engineer at sciforium

Department: Engineering

Compensation

$190,000 – $250,000/yr

Setup: On-site
Location: San Francisco, California
Level: senior
Posted: Aug 24, 2026

Description

Sciforium is hiring a Distributed Training and Inference Engineer to build, optimize, and maintain the software stack powering large-scale AI training and serving workloads. The role spans CUDA/ROCm runtimes, JAX and PyTorch frameworks, distributed system configuration, multi-node GPU cluster integration, profiling, and debugging of hardware–software interactions. Candidates should have at least five years of experience in ML systems, distributed training, or related fields, along with strong Python and C++ skills and familiarity with modern distributed training technologies.

Required Skills

Python
C++
CUDA
ROCm
JAX
PyTorch
Distributed training
Profiling
Debugging

Additional Information

Minimum Education: Bachelor’s or Master’s degree in Computer Science, Computer Engineering, Electrical Engineering, or a related technical field
Required Previous Experiences: At least 5 years of industry experience in ML systems, distributed training, or related fields
Strong programming experience in Python and C++ with familiarity with ML tooling and distributed systems
Experience with multi-node distributed training systems and orchestration frameworks such as DTensor or GSPMD
Hands-on experience maintaining or building ML training stacks involving CUDA, ROCm, NCCL, XLA, or similar technologies
Preferred Qualifications: Extensive experience with the XLA/JAX stack, including compilation internals and custom lowering paths
Familiarity with distributed serving or large-scale inference frameworks such as vLLM, TensorRT, or FasterTransformer
Background in GPU kernel optimization or accelerator-aware model partitioning
Strong understanding of low-level C++ building blocks used in ML frameworks, including XLA, CUDA kernels, and custom operations