AI Enablement Engineer

Senior AI Agent & Platform Engineer Location: Remote (LATAM preferred) Team: AI Enablement / Platform Engineering Client: Client-Facing (Embedded in Client Teams) Seniority: Senior (7–10 years experience) Summary You'll design, build, and deploy the AI agent infrastructure that transforms how engineering teams ship software. Embedded directly inside client organizations, you'll create the tooling layer that makes agentic development real — from CLI orchestrators and MCP servers to specialist agents that handle code generation, PR review, and automated QA. This isn't research or prototyping. You're building production systems that autonomous agents run on, inside real teams, under real constraints. Requirements (Must-Haves) Excellent English communication skills — you'll interact directly with client engineering leadership, present architectural decisions, and operate autonomously inside their org. If you can't communicate clearly in English, this won't work. 7+ years in software engineering with deep, demonstrable experience in backend systems, scripting, and CLI tooling. You've built things that other engineers depend on daily. Hands-on experience building AI agent systems in 2024/2025 — not just using Copilot or ChatGPT. You've built or contributed to agent orchestration, tool-use pipelines, MCP servers, Claude Skills/Commands, AGENTS.md configurations, or equivalent. Strong proficiency with Claude Code, Cursor, Codex, Copilot, or similar AI-native development tools. You use these daily and have opinions about their tradeoffs. Prompt engineering as a core technical discipline — this isn't optional or adjacent to the role, it is the role. You'll spend a good portion of your time writing structured system prompts, designing tool schemas, crafting AGENTS.md files, and building evaluation harnesses for non-deterministic outputs. The quality of every agent you build depends on the precision of your prompts. If you think prompt engineering is "just asking ChatGPT questions," this isn't the right fit. Solid understanding of cloud infrastructure (AWS preferred), containerization (Docker), and CI/CD pipelines. You'll deploy agents into real infrastructure, not notebooks. Experience building observability for software systems — logging, monitoring, alerting. You understand that agent observability (token consumption, latency, output consistency) is a new problem space you'll need to define. Comfortable operating autonomously inside a client's engineering organization — navigating their codebase, earning trust with their senior engineers, and delivering without hand-holding from Nimble. Requirements (Nice-to-Haves) Experience with agent orchestration frameworks: LangGraph, CrewAI, OpenAI Swarm/Symphony, or custom orchestration systems. Familiarity with deployment platforms for agents like Exe.dev, Modal, Harness, or similar — or you've built your own deployment pipelines for agent workloads. Background in Developer Experience (DevEx) or Platform Engineering — you've built internal tooling, developer portals, or workflow automation that made engineering teams measurably faster. Experience building or integrating with SDLC tools programmatically: Jira, Linear, GitHub Actions, QA reporting systems, or similar. Understanding of cost modeling and token economics for LLM-powered systems — you think about when to use expensive models vs. cheap ones, caching strategies, and retrieval vs. context stuffing. Security mindset for autonomous systems — you've thought about agent permissions, sandboxing, guardrails, and human-in-the-loop checkpoints. Spanish language skills (makes internal Nimble coordination significantly easier). Bonus Points Public GitHub/portfolio with agent experiments, MCP servers, CLI tools, or open-source contributions to agentic tooling — not tutorials, actual tools you built and use. You've studied or replicated architectures from: Stripe Minions, Ramp's coding agents on Modal, Coinbase's enterprise agent patterns, or OpenSWE-bench approaches. Experience with evaluation frameworks for AI systems (evals, benchmarks, regression testing for non-deterministic outputs). You've used or contributed to tools like OpenClaw, SWE-agent, or similar agent benchmarking/execution tools. Conference talks, blog posts, or workshops on agentic development, AI-assisted engineering, or DevEx.