Hi folks sharing my interview experience.

Round 1: Technical Deep Dive (RAG + Agentic AI)
This round focused on a deep discussion of my financial RAG project, covering multi-hop queries, cross-document retrieval, table handling, and Graph RAG architecture.
There were also questions on making retrieval adaptive based on query complexity.
A scenario-based discussion on agentic AI exploring agent implementation and HITL. Additionally, I was asked to implement an async programming problem. The round also covered evaluation strategies for both retrieval (recall, precision, MRR) and generation (faithfulness, grounding).

Round 2: Hands-on + System Design
This round involved building a FastAPI service with proper request validation using Pydantic, along with implementing retry logic, timeout handling, and robust error handling.
There was discussion on evaluation and experiment tracking using MLflow. Cloud-related questions included differences between serverless and AKS deployments. I was also asked about LangGraph concepts such as ReAct and reducer functions, along with observability practices using LangSmith.

Round 3: Managerial Round
This round focused on discussion into my past projects and overall contributions, with cloud deployment experience, and how I approached evaluation and observability in production systems.

HR Round
The HR round covered compensation expectations, role expectations, and a general discussion around culture and team fit.