Adobe AI Research Scientist interviews, as candidates describe them, center on a deep discussion of your past research, often a talk or project walkthrough. Some also add light probability, coding or ML questions.
$175K
Avg. Base Comp
$320K
Avg. Total Comp
Not reported
Typical Rounds
Not reported
Process Length
The clearest pattern across candidate reports is that Adobe research interviews revolve around your own work. Candidates describe walking through past projects, explaining what was novel, comparing their approach with existing methods, justifying methodological choices and saying what they hoped to explore next. Interviewers are described as collaborative but willing to push back on claims, so expect to defend your research contributions clearly.
Formats vary. A research scientist candidate described a single conversational round that included reviewing a paper the team was working on, two straightforward algorithm questions and one probability question. Another candidate presented recent video-transformer research with slides and was asked about long-range dependencies, why Adobe, and future focus areas. That conversation wrapped up in about 15 minutes. Intern reports mention an hour with a team scientist that included basic computer vision questions and probability puzzles such as an amoeba survival problem and the semicircle problem.
Coding expectations vary by role and report. Two intern candidates said there was no coding. One research associate had a Python assessment followed later by PyTorch and Python questions. Outcomes in the reports were mixed, so prepare for the format your recruiter describes.
To prepare, rehearse a project narrative that covers novelty, comparison with existing methods, trade-offs and next steps. Be ready to say why Adobe's research interests fit yours. Keep probability and basic ML fundamentals fresh in case a short check appears.
Synthesized from 10 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 Adobe process.
My interview was a single consolidated round that felt lighter in scope than I'd expected, but broader in what they assessed. I walked in knowing they'd want to hear about my research interests and prior work, so I came prepared with concrete examples. Early on, they asked me to discuss my research focus areas and what drew me to those problems. The conversation was genuinely conversational—not an interrogation, more like a team member trying to understand how I think.
They also asked me to review a paper they were already working on and give my thoughts. That was interesting; it tested whether I could engage critically with their work and see myself as part of that research direction. Beyond the research talk, there were two algorithmic questions—nothing exotic, just DSA concepts at a straightforward level—and a probability question. Nothing branch-specific or overly tricky. The whole vibe was collaborative, like they were evaluating whether I'd mesh with the team and contribute meaningfully.
One thing I'd note: the professionalism of the interview setting mattered to me. My experience felt integrated into the team from the start. The interviewers knew my background, asked thoughtful follow-ups, and signaled that research contributions were valued. I got an offer, and looking back, it felt like the right fit because the interview itself demonstrated how the team would actually treat you once you're on board.
Prep tip from this candidate
Have a clear 5-10 minute narrative of your ML research interests and a recent project ready to discuss conversationally; practice two DSA questions and one probability problem at easy-to-medium difficulty; and don't hesitate to ask clarifying questions about their research direction during the interview.
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 Adobe
Write a function to group timestamps into 7-day lists starting from the first timestamp
| Question | |
|---|---|
| Bank Fraud Model | |
| Hurdles In Data Projects | |
| Replace Words with Stems | |
| Success Measurement | |
| Testing Price Increase | |
| Z and t-Tests | |
| Overfit Avoidance | |
| RAG Strict Source Control | |
| Data Preparation for Imbalanced Data | |
| Text Editor With OOP | |
| The Longest Journey | |
| Ticket Reservation Locking | |
| Confidence Interval Explanation | |
| Decreasing Subsequent Values | |
| Proof k-Means Converges | |
| Shortest Path Algorithms | |
| 2nd Highest Salary | |
| Merge Sorted Lists | |
| Experiment Validity | |
| Employee Salaries (ETL Error) | |
| Bagging vs Boosting | |
| Scrambled Tickets | |
| First to Six | |
| P-value to a Layman | |
| Prime to N | |
| String Shift | |
| 500 Cards | |
| Compute Deviation | |
| Nearest Common Ancestor |
Synthesized from candidate reports. Individual experiences may vary.
One candidate with a referral was fast-tracked to a recruiter phone call that covered context and logistics before the technical interview. Details likely vary by team and route.
Most candidates describe the core conversation as a walkthrough of past research. Interviewers reportedly push on novelty, comparisons with existing methods, methodology and impact. Formats range from a roughly 30-40 minute discussion to a slide presentation of recent work.
Some candidates report short foundational checks after the project discussion. Examples include basic computer vision, probability puzzles, reviewing a paper, two simple algorithm questions, or PyTorch and Python questions. Two intern candidates say there was no coding at all, so this part may depend on role and team.
Some candidates say they were asked why they wanted Adobe, what research areas they wanted to pursue and how their interests match the team. Interviewers are typically described as collaborative, though some pushed back on technical claims. Prepare a clear research vision.
