We’re hiring a contract AI engineer with a machine learning background who builds, deploys and operates AI agents in production on top of large datasets, using leading-edge agentic development practice to ship quickly. Our client builds AI agents and products for the trucking and fleet compliance space, used by both its customers and its internal teams. Agents work across tolling and millions of compliance documents, on top of more than 10 years of toll, inspection, safety and vehicle data. What you ship affects revenue, safety and compliance directly.

Responsibilities:

• Design, build and deploy AI agents end to end, from first spec to production.
• Ship early and iterate on feedback from real users.
• Work with product, support and engineering stakeholders to turn loose problems into clear requirements.
• Build agents and AI features on top of the client’s data in Databricks.
• Own what you ship once it’s live: monitoring, evals, cost and failure handling.
• Help the team move faster by bringing in and spreading best practices in agentic engineering.

How you’ll work:

• Use coding agents as your main way of building software: define the work, check what agents produce and make the calls that matter.
• Break work into pieces an agent can finish and you can verify, and run several agents in parallel.
• Rely on tests, CI and agent review so speed doesn’t cost quality.
• When something goes wrong, improve your setup so it doesn’t happen again, and package what works as reusable skills for the team.

Requirements:

• Machine learning or AI engineering background.
• A track record of shipping agents and AI features to production, with users who rely on them today.
• Hands-on experience with Databricks, developing and managing batch and streaming pipelines.
• Cloud experience in Azure or AWS.
• You use evals and data to decide when an agent is ready to ship and to measure whether it’s succeeding once live.
• Strong Python, with experience building with LLM APIs and MCP.
• Experience with agent frameworks or SDKs, and knowing when not to use one.
• You already use coding agents as your main way of building software.
• You work independently, gather requirements from non-engineering stakeholders and give specific answers. When you don’t know something, you say so.
• Advanced English, written and oral.

Nice to have:

• Owning deployment of what you build: CI/CD, infrastructure as code (Terraform) and monitoring.
• Processing complex documents with LLMs (OCR, extraction).
• MLOps tools such as MLflow.
• Background in logistics, transportation or compliance.

Important:  

Remote, LATAM · Overlap with 9am–5pm US Eastern

Interviews are hands-on: expect to walk through something you’ve built and shipped, in detail.