AI Engineer - Generative AI and Agents

About the role

Azumo, a San Francisco-based software company that provides nearshore AI engineering teams, is hiring an AI Engineer to own the production behavior of AI systems once they are live: retrieval pipelines, tool-using agents, evaluation harnesses and guardrails. The role is fully remote across Latin America, aligned to the client's working day, and focuses on reliability, measurability and cost rather than demos.

Key responsibilities

• Build retrieval systems: chunking and embedding pipelines, hybrid search, reranking and retrieval quality evaluation on pgvector, Pinecone, Qdrant, FAISS or Azure AI Search.
• Develop agentic workflows: stateful multi-step execution, tool calling, MCP servers, structured output, context-window management, deterministic fallbacks and human-in-the-loop gates.
• Create evaluation suites: test sets reflecting real decisions, model-as-judge scoring, regression tracking and error analysis.
• Implement reliability and safety: prompt-injection defense, output validation, guardrails, PII handling and graceful degradation.
• Operate production deployments: containerized services on Azure or AWS, CI/CD, observability and explicit latency, cost and token budgets.

Required profile

• 4+ years shipping production software, primarily with a modern backend language such as Python, plus testing, code review, CI/CD, Git, containers, API design and async programming.
• Demonstrable production experience with LLM-based systems: RAG, function and tool calling, structured output and prompt design.
• Experience with agent frameworks and tooling such as LangGraph, LangChain, CrewAI, MCP or native Python execution loops.
• Bachelor's degree in Computer Science, Data Science or a related field, or equivalent experience.
• C1-level English and the ability to explain technical trade-offs to a client.

Required skills

• Vector and retrieval infrastructure: pgvector, Pinecone, Qdrant, FAISS or Azure AI Search.
• Evaluation tools such as Langfuse, Ragas or LangSmith.
• Cloud deployment (Azure preferred, AWS acceptable) with Docker, CI/CD and infrastructure as code (GitHub Actions, Terraform or Bicep).
• Use of AI-assisted coding tools such as Claude Code, Cursor or GitHub Copilot.

Preferred: fine-tuning with LoRA, QLoRA or PEFT, self-hosted inference, multimodal systems, LLM security work (prompt injection, red-teaming), SOC 2 or HIPAA delivery, streaming inference, open-source contributions or technical writing.

What we offer

• 100% remote-first work across Latin America.
• Paid time off and U.S. holidays.
• AI training and certifications, mentored career development and profit sharing.
• Remuneration in U.S. dollars.