Key skills
PythonRAG designChunkingEmbeddingsHybrid searchRe-rankingRetrieval evaluationAgent orchestrationLangGraphLangChainVector searchpgvectorPineconeWeaviateFastAPIBackend designDockerKubernetesCI/CDDevOps fundamentalsPostgreSQLPulumiSalesforceHubSpotOpenAIClaudeBusiness judgmentExtreme ownershipAIRAGPostgres
About this role
CargoSprint is a company focused on deploying AI agents into business teams rather than just for engineering use. They are seeking a Staff Engineer who will work closely with business leaders to understand and automate workflows, ultimately improving operational efficiency.
Responsibilities:
- Embed with business teams to find the workflows worth automating — and the ones that aren't
- Own the agent systems layer end to end: retrieval, orchestration, tooling, and evaluation
- Build RAG pipelines that ground agents in real CargoSprint knowledge
- Design multi-step agent workflows (LangGraph or equivalent) that survive real, messy usage
- Connect agents to the systems teams actually run on — Salesforce, HubSpot, Postgres, internal APIs
- Instrument everything, so you catch degradation before a user ever feels it
- Build reusable primitives so the next agent ships faster than the last
- Travel to Guadalajara as needed — the workflows you're designing for partially live there
Requirements:
- 8+ years building systems that hold up under real production load
- Expert Python; deep, hands-on RAG design experience (chunking, embeddings, hybrid search, re-ranking, retrieval eval)
- Agent orchestration in production — LangGraph/LangChain or equivalent, with tools, memory, branching, and human-in-the-loop
- Vector search fluency (pgvector, Pinecone, Weaviate) and the instinct to build an eval harness before you trust a pipeline
- FastAPI, strong backend design, and DevOps fundamentals (Docker, Kubernetes, CI/CD)
- Extreme ownership and sharp business judgment — you design for adoption, not just correctness