DevOps (LLM-ops) Freelance

Miquido is seeking an experienced freelancer to manage production Kubernetes clusters with GPU support specialized in serving large language models (LLM). The project involves deploying and configuring infrastructure tailored for GPU workloads and LLM inference.

• Experience with Kubernetes in a production environment, including GPU workloads and scheduling
• Experience serving LLM models: vLLM or SGLang
• Understanding of CUDA and GPU drivers
• Experience with Infrastructure as Code, specifically Terraform
• Familiarity with low-latency network configuration
• Experience managing storage for large model weights
• Experience with GPU observability (logs and metrics)
• Availability for client calls during the project

• Deploy and configure production Kubernetes cluster with GPU support
• Configure GPU resource scheduling
• Deploy and configure LLM serving using vLLM or SGLang
• Configure CUDA drivers and ensure compatibility across the GPU stack
• Design and implement Infrastructure as Code using Terraform
• Configure low-latency networking for LLM workloads
• Configure storage for large model weight files
• Implement observability (logs, metrics) for the GPU layer
• Participate in regular calls with the client during the project

The controller of personal data is Miquido Spółka z ograniczoną odpowiedzialnością sp. k. in Krakow, Poland. Candidates have rights regarding their personal data under applicable laws, including access, correction, deletion, processing limitation, objection, data portability, and complaint submission. Data is processed for recruitment purposes; mandatory data must be provided as required by law, and non-provision may affect recruitment. Consent withdrawal is possible anytime. Data recipients include Just Join IT and associated processors.