Job Description

Job Purpose and Impact

• Come build the AI platform that engineers across Cargill will use every day—shipping low-code and pro-code AI agents that turn real problems into real outcomes. At Cargill, you’re not optimizing vanity metrics; you’re helping a global company that puts food on tables around the world deliver better, faster, safer decisions at massive scale. We’re looking for engineers with integrity (do the right thing when no one’s watching), hunger to learn (stay curious, test, iterate), and a builder mindset (prototype, harden, scale). You’ll do well here if you move work forward even with dependencies, communicate clearly, and ship value in increments—building strong relationships while finding practical paths around blockers.

Key Accountabilities

• AI PLATFORM ENGINEERING: Builds and operates agentic applications on an AWS-based, multi-team AI platform using strong Python (typing, packaging, async/concurrency) to deliver production-grade services and APIs.
• CI/CD & QUALITY GATES: Builds, tests and releases with automated eval checks and quality gates incrementally.
• AGENTIC WORKFLOWS: Designs structured outputs, tool calling patterns, retries/timeouts and deterministic behaviors to create reliable, secure AI agents integrated through the LLM gateway.
• RETRIEVAL & RAG: Implements retrieval strategies including chunking, embedding approaches and metadata filtering at scale.
• PLATFORM COLLABORATION & GOVERNANCE: Contributes to shared platform standards including API contracts, semantic versioning, documentation and code reviews to support multiple contributing teams.
• OBSERVABILITY & OPERATIONS: Applies telemetry-first practices (logs, metrics, traces), supports on-call readiness and writes runbooks/postmortems to ensure operational excellence.
• SECURITY & RESPONSIBLE AI: Implements secure-by-default practices including secrets management, least privilege access, safe logging and protections against prompt injection, tool abuse and data leakage.

Qualifications

• Minimum requirement of 4 years of relevant work experience in:

  • Experience with LLM observability and evaluation tooling (e.g., LangSmith).

  • Familiarity with multi-model routing, fallbacks and cost/latency trade-off strategies through an LLM gateway.

  • Experience applying MCP tool interface patterns for secure tool exposure.

  • Understanding of semantic caching and performance optimization techniques for AI workloads.

  • Experience operating within a platform model that supports multiple product teams contributing to shared repositories and release trains.

  • Exposure to infrastructure-as-code and environment promotion strategies supporting reliable AI deployments.

  • Vibecoding and code assistants.