Senior Python LLM Backend Engineer
Posted Updated
Summary
Senior backend engineer in Mississauga building production LLM agent runtimes for cloud and on-prem: owning agent loops, tool calling, RAG/retrieval, and LLM traffic routing. Core stack is Python, FastAPI-style services, LangChain/LangGraph, SQL/Postgres, and TDD.
Python and LLM application engineering define this backend role at Supra ITS. Build production agent runtimes for cloud and on-premise deployments.
You will own agent loops, tool calling, retrieval, file context, and LLM traffic routing. The work uses Python, FastAPI or similar services, LangChain or LangGraph, SQL, Postgres, and Test Driven Development. You will ship through stories, architecture decision records, acceptance criteria, review, and QA. Production behavior matters more than notebook research or prompt-only consulting.
Key Responsibilities:
Design agent loops, prompts, state, persistence, and streaming SSE Harden MCP and native tools with clear failure modes Improve chunking, embeddings, hybrid search, and grounded citations Integrate LiteLLM routing, timeouts, streaming, and attribution Keep attachments, artifacts, and knowledge stores correctly separated Requirements:
Strong Python experience in production services Hands-on function calling, RAG, prompting, and evaluation loops LangChain or LangGraph experience debugging multi-turn traces Solid SQL and Postgres knowledge, including migrations and RLS Three to five-plus years engineering, including LLM feature delivery Own reliable Python agent infrastructure through TDD, retrieval, tool safety, and production-ready LLM routing at Supra ITS.
Design agent loops, prompts, state, persistence, and streaming SSE Harden MCP and native tools with clear failure modes Improve chunking, embeddings, hybrid search, and grounded citations Integrate LiteLLM routing, timeouts, streaming, and attribution Keep attachments, artifacts, and knowledge stores correctly separated Requirements:
Strong Python experience in production services Hands-on function calling, RAG, prompting, and evaluation loops LangChain or LangGraph experience debugging multi-turn traces Solid SQL and Postgres knowledge, including migrations and RLS Three to five-plus years engineering, including LLM feature delivery Own reliable Python agent infrastructure through TDD, retrieval, tool safety, and production-ready LLM routing at Supra ITS.