Distributed training engineer
This role demands strong engineering fundamentals, deep systems understanding, and a passion for squeezing every last bit of performance from modern compute infrastructure; Collaborate with researchers to build system…
- Employer: Soniox
- Location: Slovenia
- Employment type: Full-time
- Work mode: ON_SITE
- Posted:
- Apply before:
Job description:
- At Soniox, we are building world-class speech and multimodal AI systems—and scaling their training to the limits of modern hardware. As a Distributed training engineer, you will work closely with researchers to design and optimize large-scale training infrastructure. Your work will directly impact the throughput, stability, and performance of our models, enabling rapid experimentation and unlocking new capabilities across speech, video, and language.
- This role demands strong engineering fundamentals, deep systems understanding, and a passion for squeezing every last bit of performance from modern compute infrastructure.
- Collaborate with researchers to build systems-efficient training loops for cutting-edge speech and video models.
- Profile and optimize distributed training workloads across our internal infrastructure.
- Design and implement improvements to our in-house training framework to increase hardware utilization and throughput.
- Apply the latest algorithmic and systems-level techniques to improve scalability, reproducibility, and stability in large-scale training.
- Investigate and resolve performance bottlenecks across the ML stack—from dataloaders to kernel fusion to network latency.
Requirements:
- Have experience working with large-scale or multi-modal ML pipelines.
- Enjoy diving into the internals of deep learning frameworks and distributed systems to make them faster and more reliable.
- Are proficient in Python and have strong software engineering practices.
- Understand how training kernels work and how to optimize them (e.g., CUDA, Triton, XLA, or custom ops).
- Are passionate about making ML training stable at scale, not just functional.
- Care deeply about correctness, reproducibility, and engineering craftsmanship.
We offer:
- You’ll help build one of the most technically advanced AI platforms in the world — and shape how it reaches and supports users globally.
- You’ll work directly with a world-class team of engineers and researchers solving frontier problems in speech and language AI.
- You'll have a voice in how our company grows, how our customers succeed, and how AI transforms human communication.