Oxygen X AI Engineer interview typically runs 2 rounds: HR screening, then a technical interview. The process is straightforward and leans heavily on ML concepts over behavioral questions.
$149K
Avg. Base Comp
$182K
Avg. Total Comp
2
Typical Rounds
2-4 weeks
Process Length
Our candidates report that Oxygen X moves quickly past surface-level ML knowledge and into applied reasoning — specifically around high-stakes fintech problems. The clearest signal from the one experience we have is the fraud detection and credit scoring scenario, which wasn't a warmup question but the centerpiece of the technical round. That framing tells us a lot: they want to see how you structure a problem, not just which algorithm you'd reach for. The ability to justify modeling choices in a risk-sensitive, bank-facing context appears to be the real evaluation criterion here.
What's notable is how the question mix spans both ends of the ML spectrum — from definitional questions like "what is supervised learning" to comparative judgment calls like when to use K-means over KNN. This suggests the team is checking for conceptual fluency as a baseline, then probing whether you can reason through tradeoffs. Evaluation metrics for deep learning came up explicitly, which in a fintech context almost certainly means they care about precision-recall dynamics, not just accuracy. The enterprise experience question also stood out — it hints that Oxygen X values candidates who've worked within organizational constraints, not just built models in isolation. Recovery from missed questions seems possible here, which suggests the process rewards overall coherence over perfection on any single answer.
Synthesized from 1 candidate report by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Oxygen X process.
Straightforward two-round process. The first round was a preliminary HR interview. The second round was a technical interview focused on machine learning concepts and scenario-based questions.
I was asked things like: "You are asked to build a fraud detection or credit scoring system for a bank — how would you approach it?" The questions were mostly centered on supervised and unsupervised learning techniques and evaluation metrics. There was less focus on behavioral questions and more on technical ability.
Specific questions I remember:
I missed a couple of questions but there were enough rounds that I was able to recover. Still waiting for a reply.
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 Oxygen X
Write a query that returns all neighborhoods that have 0 users.
| Question | |
|---|---|
| Rolling Bank Transactions | |
| 2nd Highest Salary | |
| Top Three Salaries | |
| Comments Histogram | |
| Closest SAT Scores | |
| Employee Salaries | |
| Cumulative Distribution | |
| Subscription Overlap | |
| Size of Joins | |
| Paired Products | |
| Hurdles In Data Projects | |
| Merge Sorted Lists | |
| Top 5 Turnover Risk | |
| Like Tracker | |
| Last Transaction | |
| Month Over Month | |
| String Shift | |
| Find the First Non-Repeating Character in a String | |
| Top 3 Users | |
| Bagging vs Boosting | |
| Total Spent on Products | |
| Third Purchase | |
| Daily Logins | |
| Alphabet Sum | |
| Get Top N Frequent Words | |
| Minimum Change | |
| Prime to N | |
| P-value to a Layman | |
| Largest Salary by Department |
Synthesized from candidate reports. Individual experiences may vary.
The process starts with a preliminary HR interview. This round is mainly an introductory screen to confirm interest in the AI Engineer role, review the candidate's background, and check general fit before moving into technical evaluation.
The second round is a technical interview centered on machine learning concepts and scenario-based problem solving. Candidates should expect questions on supervised learning, unsupervised learning, and how to choose methods such as K-means versus KNN in practical settings.
A major part of the technical conversation is an applied case, such as designing a fraud detection or credit scoring system for a bank. The interviewer probes how you would approach the problem end to end, including model choice, evaluation, and tradeoffs.
The interview also covers how to evaluate machine learning and deep learning models. Questions focus on which metrics you would use and why, with an emphasis on demonstrating practical judgment rather than memorized theory.
Candidates are asked about prior experience with enterprise companies and the most challenging project they have worked on. This portion is less behavioral overall, but it helps the interviewer understand the depth and relevance of your past work.
