Many Abnormal AI software engineer candidates describe a practical, AI-friendly process: a take-home, a follow-up on that work, system design with incident handling, and code review or manager rounds. Some candidates instead report conventional coding rounds.
$182K
Avg. Base Comp
$271K
Avg. Total Comp
4 rounds
Typical Rounds
Not reported
Process Length
Candidates describe an Abnormal AI software engineer process that leans toward practical engineering over algorithm puzzles, and AI tools come up from the first technical stage. Several reports start with a take-home where AI use was allowed or encouraged. One candidate built a file storage system, another had 24 hours to build a file storage vault, and a third had 10 days for a frontend and backend assignment. Two candidates said they recorded a video explaining how they used AI, so be ready to explain your AI workflow, not just show working code.
What follows varies by report. Some candidates describe a second round built on the take-home, where the interviewer asks about design decisions and then asks for another feature, such as an API to search files for specific words. Later rounds in different reports include system design combined with incident handling, a code review or debugging round where you fix problems like SQL injection and null references, a past project review, and an engineering manager conversation. One candidate found the AI-assisted coding setup awkward and was repeatedly nudged to keep talking, so practice narrating your thinking while working with an AI assistant.
Not every report matches that pattern. One candidate described a more conventional sequence: two coding problems, a backend and system design discussion covering APIs, caching and locking, a full-stack round, and a hiring manager conversation. Another described a recruiter and hiring manager call with behavioral questions and two coding problems. Loop structure differs across reports, so ask your recruiter what your loop looks like.
Outcomes are mixed. Most reports ended in a rejection or no offer, and two candidates were still waiting to hear back. Two candidates also mentioned weak communication after the take-home, including rejections with no feedback, so keep written confirmation of anything you submit.
Synthesized from 13 candidate reports by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Abnormal AI process.
Share your own interview experience to unlock all reports, or subscribe for full access.
Sourced from candidate reports and verified by our team.
Topics based on recent interview experiences.
Featured question at Abnormal AI
Describing a data project and its challenges
| Question | |
|---|---|
| Internet Congestion | |
| Client Solution Pushback | |
| Testing Constraints | |
| 2nd Highest Salary | |
| Empty Neighborhoods | |
| Top Three Salaries | |
| Merge Sorted Lists | |
| Largest Salary by Department | |
| Top 3 Users | |
| String Shift | |
| Find Bigrams | |
| Last Transaction | |
| Prime to N | |
| Raining in Seattle | |
| Size of Joins | |
| Minimum Change | |
| First Touch Attribution | |
| Job Recommendation | |
| Find the First Non-Repeating Character in a String | |
| Daily Retention Summary | |
| Find the Index with Equal Left and Right Sum | |
| Google Maps Improvement | |
| Delivery Estimate Model | |
| Type-ahead Search | |
| Detecting ECG Tachycardia Runs | |
| Friendship Timeline | |
| Type I and II Errors | |
| Term Frequency | |
| The Brackets Problem |
Synthesized from candidate reports. Individual experiences may vary.
Candidates describe starting with a phone or recruiter conversation. One candidate had a combined recruiter and hiring manager call. It covered behavioral questions about past projects and production incidents, followed by two coding problems.
Several candidates report a take-home with AI use permitted. Examples include a file storage system, a file storage vault with a 24-hour window, and a frontend and backend assignment with 10 days. Two candidates mention recording a video that explains their AI usage.
Some candidates describe a follow-up round based on the submission. In one report, the interviewer asked about design decisions and then requested an API to search files for specific words. Practice narrating your reasoning while you code.
Reports describe a round that combines incident handling with system design. One candidate read charts to find why a backend service was failing. They then gave feedback on an existing email-processing design rather than designing one from scratch.
One candidate reviewed code for glaring problems such as SQL injection and null references and advises not chasing nitpicks. Another had a code review and fix implementation round using an AI agent, where the interviewer kept nudging them to talk through their thinking.
Candidates report manager rounds covering how they approach problems, work with teammates, handle disagreements and learn new technologies. One candidate's loop also included a past project review. Prepare concrete project stories.
