agentic ai
Agentic AI Platform
Internal production platform · Tru
Agent workflows at Tru: retrieval, multi-step tasks, and making that hold up in production.
Software Engineer·2024 – Present·Mohali
Enterprise · NDA
Context
At Tru I work on an internal agent platform used for multi-step workflows across systems the company already had. I did not build Tru as a product; I own services, orchestration, and operator surfaces on that platform.
Problem
Teams needed agents that could retrieve context, run tasks in sequence, and call existing APIs without failing silently.
- —Production stability and observability
- —Retrieval latency versus answer quality
- —Agent state across partial failures
What I worked on
- →Implemented FastAPI services that host agent workflows
- →Wired Weaviate retrieval and PostgreSQL for durable agent state
- →Built execution graphs with LangGraph and LangChain
- →Shipped React surfaces for monitoring and operator interaction
Architecture & approach
Layered services: FastAPI at the edge, LangGraph for multi-step execution, Weaviate for retrieval, PostgreSQL for state, React for operators.
Outcome
- ✓Production agent workflows with retrieved context
- ✓Less manual handoff on sequenced internal tasks
- ✓Clearer failure paths through debugging and review
Stack
FastAPIPythonLangGraphLangChainWeaviatePostgreSQLReact