Senior Developer Infrastructure Platform

About the role

We're building the platform that other engineers build on top of. You'll design and ship production services, internal SDKs, and Kubernetes operators that abstract away the complexity of our hybrid infrastructure; spanning CloudStack, bare-metal GPU clusters.

Responsibilities

• Design, implement, test, and maintain production services, internal SDKs, and Kubernetes operators.
• Own services end-to-end, including debugging critical path issues when they arise.
• Write code that other teams depend on, focusing on API contracts, backward compatibility, and not breaking things quietly.

Requirements

• Java; not just familiarity. Understand concurrency, JVM internals, and work in large codebases with real engineering process.
• Golang; production services, not just tooling scripts. Write Go that other engineers depend on.
• Python; clean, testable, packaged. Not ad-hoc scripts.
• Kubernetes; deep. Write an operator or custom controller from scratch, not just apply Helm charts.
• OpenStack or CloudStack; you can explain how a compute scheduler works, not just 'we used it.'
• gRPC / protobuf; service design, streaming, schema evolution.
• API design discipline; versioning, backward compatibility, contracts. Think about what happens when you break something downstream.
• Hardware awareness; debug at the hypervisor, driver, or bare-metal level at least once.

Strong Plus

• GPU compute; CUDA, MIG, vGPU, device plugins, PCIe passthrough, SR-IOV.
• Agentic AI / ML (practitioner level); LLM API integration, agent frameworks (LangChain / LlamaIndex), RAG pipeline design, vector DBs, MLOps fundamentals. Ship a system where a model is in the critical path and debug it when it misbehaves.
• Inference basics; vLLM, Triton, or similar. Enough to understand what the platform needs to support.
• Open Source development experience, on a large project.

What We Don't Want

• A DevOps engineer who primarily operates infrastructure and writes pipelines. This role ships software.
• Someone whose Java experience stops at Spring Boot CRUD applications.
• Someone who can't discuss software design trade-offs under interview pressure; not just 'what we did' but 'what we considered and why.'