AI Platform Engineer

Nexus Weather and Climate is hiring an AI Platform Engineer to build and operate the platform infrastructure that connects its AI agents to live data, external systems, and customer workflows. The role spans MCP servers, APIs, agent and RAG systems, Python backend services, and platform reliability.

Responsibilities:

• MCP Servers & APIs: Build MCP servers that give agents live access to data and systems. Design REST API endpoints, gateway layers, and versioned contracts across the stack
• Routing Logic: Determine when to leverage RAG, MCP, or fine-tuned models to optimize AI agent performance, accuracy, and the quality of customer-facing responses
• AI Agent Development: Design, build, and deploy production-ready AI agents that integrate with Nexus data, APIs, and external systems to deliver reliable, customer-facing functionality
• Quality & Safety: Build automated evaluation harnesses at customer-safe thresholds and implement human-in-the-loop checkpoints for high-stakes agent decisions
• End-to-End RAG: Build and maintain RAG pipelines: semantic chunking, embedding model selection, hybrid search, indexing, and automated ingestion as content changes
• Vector Store Quality: Monitor retrieval accuracy, tune chunking and embedding parameters, and maintain index quality to keep AI output grounded and low-latency
• AI Backend Development: Build and maintain Python backend services on GCP that support AI agents, tool integrations, and customer-facing applications
• APIs & Webhooks: Develop and integrate REST APIs and external webhooks that allow AI agents to access live data, interact with external systems, and trigger customer workflows
• Data Integration: Connect AI agents to existing Nexus data sources and databases, including PostgreSQL, ensuring reliable and secure access to the information they need
• CI/CD & IaC: Apply DevOps practices using Terraform, Kubernetes, and GitHub Actions with AI evaluation harnesses embedded in deployment so quality regressions are caught before production
• Platform Monitoring: Monitor MCP server and AI platform uptime, performance, and costs. Respond to degradation proactively. Use GitHub Copilot and Cursor natively as the primary development mode
• Code Review & Unblocking: Review pull requests across the full stack and float into frontend, backend, AI, or infrastructure work to pair on hard problems and unblock the team
• Standards & Architecture: Define coding standards, API conventions, and architectural patterns. Surface technical debt and architectural risks without waiting to be asked