Summary
Selected for a Software Development Engineer role at Intuit through a referral. Interview process included a TPI phone round with DSA questions and system design discussions, followed by an onsite loop with four rounds covering coding, system design, AI concepts, and a hiring manager interview.
Full Experience
Profile: Software Development Engineer at a MAANG company with 2.3 years of experience. I applied through a referral from a LinkedIn connection, and the recruiter reached out within two days to schedule the initial interview.
Round 0 (TPI Phone Interview):
It was a 90 minutes round consisting of DSA and discussions on system designing. After a quick introduction recruiter asked 2 DSA questions:
- Plus One problem with some variations (https://leetcode.com/problems/plus-one/description/)
- Validate IP Address (https://leetcode.com/problems/validate-ip-address/).
After solving this there were some general questions on high level designs of a system and AI - how to ensure an API has no downtime, what models are used in trending AI apps, etc. It was a pretty fun, interesting discussion and the interviewer was very active and engaging.
The round went well, and within a few days, the recruiter scheduled my A4A onsite loop, which consisted of four rounds to be completed on the same day. I was also asked to prepare a PPT for my introduction and set up a given code repository beforehand.
Onsite – 4 Rounds (Same Day)
Round 1 – Craft Demonstration (90 mins)
Started with my introduction and projects from the PPT I had prepared. I shared and after some follow up question on experience they gave me the task.
Task was simple, just need to add another API in the given repo. I discussed my approach and panel grilled a little on the system design aspects. After I completed the basic code there was some discussion around the edge cases and how to handle errors/exceptions. I tested the new APIs on Postman and they seemed to be working as expected. Then there was some discussions on how to improve the efficiency of the functions etc. The entire code was in Python, using Flask for http routings and Docker for containerization of a localised LLM.
I hadn't coded in Python in some time(the only options for the repo are in Python & Java and I've been a C-based coder mostly) so was a little nervous in this round. I definitely believe I could have completed the coding faster had I been working in Python more recently but atleast I completed and managed to run it end-to-end which was good enough I guess.
Round 2 – Assessor (60 mins)
Two interviewers from the first round joined again.
First discussion was around database design for my code earlier. How to add some other entities in the database, defining their relationships to the existing entities. The dB being used in the code was relational so I framed my answers around SQL databases and what joins and normalisation I would use to create and connect new entities. Interviewer asked me to write code for some of the SQL commands I had suggested. Then he asked me how these tables and joins could be transformed into a graph database. I gave him a basic schema of the graph database and node structure I could think of and he seemed satisfied.
The remaining 20 minutes I was asked to solve a given problem on Glider. Question was similar to this- https://leetcode.com/problems/edit-distance/description/. I gave a brute force then optimised dp solution. Thankfully could choose any language to code in this round so back to trusty C++ it was :). Ran and tested the code, seemed to be working fine. The interviewers seemed happy with the solution.
Round 3 – AI Round (30 mins)
This round was more conceptual.
It started with unit testing—I was asked what tests I would write for the API from Round 1. I explained sample positive and negative cases, E2E testing approaches, and monitoring metrics used in production.
The discussion then moved to AI concepts such as- what is an RAG, what are system prompts, what does temperature mean in LLMs?
Round 4 – Hiring Manager (30 mins)
This round for me was a little hectic since the HM seemed to be quite busy that day. This was rescheduled almost 3 times before she could finally join. She asked me about my previous rounds, previous company and the project I had mentioned in my introductory PPT. After some follow-ups on these topics it was generalised HM questions- how do I manage work pressure, why looking for switch etc.
The conversation was short and to the point, and it was hard to read the HM’s reaction.
The recruiter reached out a few days later and confirmed that the overall feedback was positive.
🟢 Verdict: Selected