Full-Stack Backend Engineer

About the role

We are looking for a Full‑Stack Backend Engineer to own the AI and data backbone of COSO’s construction intelligence platform. You will design, build, and operate ingestion pipelines, embeddings, retrieval, and agentic orchestration that power real‑time decision‑making on large infrastructure projects.

Key responsibilities

• Collaborate with the Head of AI Engineering and the founder on architecture and customer‑facing deployments.
• Build and maintain ingestion pipelines for multi‑format construction data.
• Develop and operate embeddings and retrieval infrastructure, including vector databases and hybrid search.
• Implement the scheduling intelligence engine and agentic orchestration using the Anthropic SDK.
• Create evaluation and observability layers: golden sets, regression tests, latency and cost dashboards, and full traceability of agent runs.
• Design knowledge‑retention strategies that keep learned patterns durable and contractually defensible.
• Deploy production services with Docker, CI/CD, PostgreSQL, Redis, object storage, and task queues.
• Ensure citation‑verified outputs and monitor statistical health of models.

Required profile

• Computer Science degree or equivalent practical experience.
• 2‑5 years of backend, production AI/ML, data engineering, or applied data science experience, with ownership of at least one end‑to‑end production system.
• Strong understanding of statistical analysis and ability to validate real‑world data.
• Experience being on‑call and handling production incidents.

Required skills

• Python (async patterns), FastAPI, Pydantic.
• PostgreSQL, Redis, object storage, task queue systems.
• Docker and CI/CD pipelines.
• Anthropic and OpenAI SDKs, prompt engineering, function calling.
• Vector databases, hybrid search, reranking, query rewriting.