Principal AI Inference Platform Engineer

Optomi; Plano (TX); On-site; USD 170,000 - 250,000; Full time

Job summary

Optomi, in partnership with a large telecommunications company, seeks a Principal Inference Software Engineer to join their Plano, TX team. You will build, deploy, and scale production AI inference infrastructure across Kubernetes, vLLM, Python, GPUs, and cloud/on-prem environments.

In this role you will optimize model serving, manage GPU resources, and develop automation to support both R&D and production workloads. A strong background in distributed systems and reliability is essential.

Qualifications

• Strong software engineering and Python skills.
• Production experience with LLM inference/model serving, ideally vLLM.
• Experience with GPU infrastructure, containers, and CI/CD.
• Understanding of scaling, automation, observability, and reliability for distributed systems.
• Experience with cloud and/or on-prem infrastructure preferred.

Responsibilities

• Build and operate LLM inference infrastructure at scale.
• Deploy models with vLLM and Kubernetes.
• Develop Python automation for deployment, scaling, updates, rollbacks, and health monitoring.
• Manage GPU capacity, performance, reliability, and production workloads.
• Build tools to simplify model deployment for engineers.
• Support R&D and production inference environments.
• Troubleshoot and optimize model-serving performance.

Skills

Python

Software engineering

LLM inference

Model serving

vLLM

GPU infrastructure

Containers

CI/CD

Observability

Distributed systems

Cloud/On-prem

Tools

Kubernetes

Docker

CI/CD tooling

Job description

Optomi, in partnership with a large telecommunications company, seeks a Principal Inference Software Engineer to join their Plano, TX team. You will build, deploy, and scale production AI inference infrastructure across Kubernetes, vLLM, Python, GPUs, and cloud/on-prem environments.

In this role you will optimize model serving, manage GPU resources, and develop automation to support both R&D and production workloads. A strong background in distributed systems and reliability is essential.