Training Infra Engineer for Scalable LLM Systems

Inception; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Inception is seeking engineers and scientists to design, optimize, and maintain the core systems that enable scalable, efficient training of LLMs. Your work will make experimentation and training fast and reliable so the team can focus on science, not infrastructure bottlenecks.

You'll design distributed training systems across thousands of GPUs, develop high‑performance optimizations, and build reusable frameworks to improve training reproducibility and scalability for new model architectures.

Qualifications

• BS/MS/PhD in Computer Science, Engineering, or a related field (or equivalent experience).
• Understanding of ML frameworks (PyTorch, TensorFlow) from a systems perspective.
• Strong engineering skills — ability to contribute performant, maintainable code and debug in complex codebases.
• Proficiency in Python and at least one systems programming language (C++/Rust/Go).
• Experience with containerization (Docker), orchestration (Kubernetes), and CI/CD pipelines.

Responsibilities

• Design, implement, and optimize distributed training systems that scale across thousands of GPUs and nodes.
• Develop high-performance optimizations to maximize throughput and efficiency.
• Develop reusable frameworks and libraries to improve training reproducibility, reliability, and scalability for new model architectures.

Skills

Distributed systems

High performance optimization

Python

C++/Rust/Go

Docker

Kubernetes

ML frameworks: PyTorch, TensorFlow

CI/CD pipelines

Experimentation tooling

Education

BS/MS/PhD in Computer Science/Engineering or related field

Tools

Docker

Kubernetes

Kubeflow

Airflow

Prometheus

Grafana

OpenTelemetry

PyTorch

TensorFlow

PyTorch/XLA

DeepSpeed

Megatron-LM

Job description

Inception is seeking engineers and scientists to design, optimize, and maintain the core systems that enable scalable, efficient training of LLMs. Your work will make experimentation and training fast and reliable so the team can focus on science, not infrastructure bottlenecks.

You'll design distributed training systems across thousands of GPUs, develop high‑performance optimizations, and build reusable frameworks to improve training reproducibility and scalability for new model architectures.