All work

03 / RAG & agentic AI

Bringing AI into production systems.

RAG and agentic workflows, with evaluation and observability built into the integration.

ROLE
Lead Software Engineer · Operanex
ROLE PERIOD
Current role · Jan 2022 – Present

Three parts of production AI

RetrievalWorkflowsEvaluation
Conceptual view of the integration’s capabilities, not a sequential runtime pipeline or production topology.

Context

My recent work extends backend and real-time platform engineering into AI-enabled products, with production integration of RAG and agentic AI capabilities.

My responsibility

I lead the production integration of RAG and agentic AI using LangGraph and LangChain, including retrieval pipelines and LangSmith-based evaluation and observability.

Engineering approach

  • Develop retrieval pipelines as part of the RAG integration.
  • Use LangGraph and LangChain to deliver agentic AI capabilities and workflows.
  • Use LangSmith for evaluation and observability, with reliability, scalability, and business value as engineering priorities.

Outcome

Delivered RAG and agentic AI capabilities as part of production integration, including retrieval pipelines and LangSmith-based evaluation and observability.

Discuss my experience Next: Backend & real-time data