Abnormal AI | Software Engineer II | Interview Experience
Sharing my recent interview experience with Abnormal AI in case it helps anyone preparing for the process. Overall, it was a good experience, and the recruiter was communicative and helpful throughout.
Interview Process
The process consisted of the following rounds:
1. Recruiter Screen
The first round was a typical recruiter conversation.
We discussed:
- My previous experience
- Projects I had worked on
- The biggest impact I had created
- My responsibilities in previous roles
- What the interview process would look like
The recruiter also explained that there would be a screening round followed by several onsite/technical rounds.
2. Live Coding / Codebase Walkthrough
This was a little different from a traditional LeetCode-style coding round.
I was given an existing codebase and was expected to:
- Go through the codebase and explain how it worked.
- Understand the existing architecture/flow.
- Implement an additional feature based on a problem statement provided by the interviewer.
- Discuss possible extensions and modifications to the implementation.
There were also several follow-up questions around the implementation and how I would modify the solution under different requirements.
The extensions were mostly discussion-based; I wasn't required to implement every extension.
Preparation tip: Don't prepare only for DSA. Being comfortable reading an unfamiliar codebase and reasoning about existing code is important for this round.
3. Incident Reporting + System Design
This round had two parts.
Incident Reporting
I was given access to an AWS environment and had to investigate an incident.
The goal was to figure out:
- What was going wrong
- How I would identify the root cause
- What signals/logs/monitoring tools I would look at
- What the immediate mitigation should be
- What long-term fixes I would recommend
The interviewer was interested not just in finding the issue, but also in how I approached debugging and how I prioritized short-term mitigation vs. long-term remediation.
System Design
The second part was a system design exercise.
I was given an existing system and asked how I would scale it.
The design itself wasn't extremely complicated. The main focus was around:
- Scaling reads
- Scaling writes
- Identifying bottlenecks
- Thinking about how the existing architecture would behave as traffic increased
4. Code Review + System Extensions
This was probably one of the more challenging rounds.
I was given a small repository and asked to perform a code review.
The expectation was to:
- Understand the code
- Identify issues
- Prioritize them (P0/P1/etc.)
- Explain why certain issues were more important
- Suggest concrete improvements
The second part involved extending the system and discussing how it would behave at larger scale.
This discussion went fairly deep into:
- Concurrency
- Parallelism
- Worker scaling
- Message queue scaling
- Throughput
- Bottlenecks
- Failure scenarios
The interviewer pushed quite a bit on how the system would behave under scale and concurrency, rather than just asking for high-level architectural components.
Preparation tip: For senior-level roles, be prepared to go beyond "add more workers" or "use a queue." You should be able to reason about the limits of those approaches and explain what happens when individual components become bottlenecks.
5. Managerial / Behavioral Round
This was primarily focused on my previous work and behavioral questions.
We discussed:
- Past projects
- Technical decisions I had made
- Challenges I had faced
- Impact of my work
- General behavioral/situational questions
This was initially supposed to be the final round.
6. Technical Deep Dive — Additional Round
After completing the above rounds, I was asked to take one additional technical round because the team wanted some additional signal.
This was more of a technical deep dive into my experience and decision-making.
Although I had cleared the official interview rounds, I ultimately wasn't able to clear this final additional round.
The feedback was that the interviewer felt I lacked sufficient judgment in a few areas.
My profile was subsequently put on hold / the process did not move forward. I'm not completely sure whether it was formally a rejection or a hold, but we did not proceed further.
Overall Experience
Despite the outcome, my overall experience with Abnormal AI was positive.
The recruiter was quite communicative throughout the process, explained the different stages, provided context and tips, and generally kept me informed.
Hope this helps someone preparing for the process!