Текст заказа
About PolarGrid
PolarGrid is an early-stage Canadian company that runs managed AI inference. A customer brings a model or a serving image. We run it on dedicated GPUs behind one endpoint, serve each request from the nearest site, and move traffic to the next nearest site when one fails.
Our goal is to make AI inference fast, reliable, and secure for customers running production workloads.
The Role
We're looking for a Backend Engineer to help build, operate, and scale the infrastructure behind PolarGrid's inference network.
You'll work across on-premises and cloud distributed services, infrastructure, reliability, and security. This is a hands-on engineering role where you'll be responsible for building, running, and supporting systems in production.
You'll work closely with a small engineering team and take responsibility for technical decisions, implementation, and operational reliability.
What You'll Do
Build and operate distributed compute services, control plane, gateway, and telemetry agents on a fleet of inference GPU servers.
Build and operate the DNS and networking layer that route customer requests between sites.
Automate deployment of customer model images on our own GPU servers.
Write infrastructure as code, ship versioned releases through CI, and set up monitoring and alerting.
Apply security best practices to authentication, data, secrets management, and tenant isolation.
What You've Done
Built and run production backend services in Python, Go, or Rust, including on-call.
Set up and run open-source models on inference backends, such as vLLM, SGLang, TensorRT-LLM, llama.cpp, or Triton Inference Server.
Operated bare-metal Linux servers and containerized workloads.
Written infrastructure as code and CI pipelines.
Implemented DNS or load balancing.
Applied security best practices and test-driven development.
Originally posted on Himalayas