Tiger Analytics AI Engineer Interview Guide

I interviewed at Tiger Analytics for an AI Engineer role across three technical rounds spanning almost three hours. The first round was an AI-proctored interview that covered foundational coding—I got a factorial problem and a 3-sum problem, both standard DSA fare. The second round was a one-on-one virtual call where things shifted toward machine learning and deep learning concepts. We dove into RAG architecture, agentic AI systems, deployment strategies, and cloud platforms like AWS and Azure. The interviewer also asked behavioral questions, including why I wanted to make a move. By the third round, which was in-person, the focus became even more specialized. We discussed RAG pipelines in depth—preprocessing, chunking strategies, vector databases, retrieval flows, and how to evaluate and troubleshoot these systems in production. One question that stuck with me was asking me to walk through what happens internally when a user asks a chatbot a question, step by step. The interviewers were polite and engaged throughout, and I felt I answered most questions clearly and demonstrated solid understanding of the material. What caught me off guard was the outcome. I was told afterward that I wasn't being shortlisted, but I received no feedback explaining why or what I could improve. That lack of closure made the whole experience feel incomplete, even though the technical discussions themselves were substantive and thoughtful.