real interview stories, told over a cup of tea ♡
Interview Experiences
3935 interview stories from candidates at 857 companies: how each round went, what was asked and how it ended. Every company page links to practice questions in the same style.
- 3935
- Experiences
- 857
- Companies
- 532
- Offers
3935 results
- ML Fundamentals for a Research RoleMistral AI · Software Engineer · Technical Screen · Sep 2026Rejected
The author reports an unsuccessful interview for a research-engineering role at Mistral. After an introduction to the team, the applicant spent several minutes discussing relevant background and projects. The interview then moved through direct questions about model parallelism, mixtures of experts, normalization, diffusion, optimization, preference-based training, and keeping GPUs supplied with…
- ML Fundamentals Plus a Production PR Review, RejectedNetflix · Machine Learning Engineer · Technical Screen · Sep 2026Rejected
Sharing a failure story from September. The recruiter reached out to me on LinkedIn, and then we scheduled a phone screen. Since this was hiring for a newly formed applied GenAI org, the format was pretty new and I haven't seen it in any other interview write-ups: 30 min of ML fundamentals + 30 min of practical coding. 30 min ML fundamentals, three questions: What is a KV cache? What is attention…
- ML Multiple Choice and Production Request SchedulingC3 AI · Data Scientist · Intern · Online Assessment · Sep 2026
Question 1: Vanishing Gradient Which of the following is NOT a potential solution to the vanishing gradient problem in a deep neural network? Adding skip connections to the neural network architecture Using a different activation function like ReLU Raising the learning rate Adding Batch Normalization between layers Question 2: Boosting Classification For the classification problem, Boosting is wi…
- Model Deep Dives and Organization-Chart CodingSnapchat · Machine Learning Engineer · Onsite · Sep 2026
The author describes Snapchat onsite sessions divided between a short values-and-project discussion and approximately 50 minutes on the round’s technical subject. Values questions emphasized AI applications, prioritization, and empathy. The machine-learning discussion examined project decisions and training problems, including normalization, regularization, loss functions, and different predictio…
- Mouse, Cheese and Cat Grid Path, a Googlyness Manager Question, and a Non-Axis-Aligned Max RectangleGoogle · Software Engineer · Onsite · Sep 2026
Phone Interview 1 Warmup Question: Given an N x N grid with a start cell S and a target cell T, where some cells are impassable water. You can only move horizontally or vertically. Determine whether a path from S to T exists. Main Question: Given an N x N grid. A mouse starts at S and the goal is the cheese at T. The grid also has impassable water cells and a cat C. The mouse can only move horizo…
- Music Player Design and Product DiscussionsAtlassian · Software Engineer · Senior+ · Onsite · Sep 2026
The author describes interviewing for an ML/AI product engineering role with BCNY, a small company acquired by Atlassian, and cautions that this team’s process may differ. A phone screen used a music-player exercise to assess object-oriented design and communication, including playback controls, queuing, repeat behavior, and shuffled playlists. Onsite discussions included designing comments for a…
- New Card Game Coding Round That Tests Extensible OOD DesignOpenAI · Research Scientist · Onsite · Sep 2026
The role was RS. I got a new coding problem, a card game. The first three parts were: first, implement a comparison where three of a kind > pair > high card; second, add a new hand so the order becomes three of a kind > flush > pair > high card; third, a discard rule (the rule I got is not quite the same as the one people posted on the forum, mine was to discard if it is not a pair). You can use…
- OA on the Longest Risk-Limited Trade-Log WindowFalconX · Software Engineer · Online Assessment · Sep 2026
Build an audit module for a high-frequency trading platform that processes an asynchronous, out-of-order data stream containing trade execution logs from multiple exchanges. The task is to reconstruct the chronological execution timeline for a specified exchange (target_venue), and find the maximum number of consecutive trade logs in a contiguous execution window that satisfies specific operation…
- OA with the Deployment Window Scheduler, Part 2 UnfinishedStripe · Software Engineer · Online Assessment · Sep 2026
This was the OA round. It's the Deployment Window Scheduler problem that has already been posted on the forum. Build a deployment-window scheduler over a recurring 10,080-minute week. Find all deployable time slots. Part 1: the input is , where type is either allowed or freeze. Output the deployable time slots . Part 2: the first line of the input is…
- OA, a Great Recruiter, and a Tree-Rendering VO Where I Finished Two PartsStripe · Software Engineer · Technical Screen · Sep 2026In progress
I got the OA the same day I applied, probably a mass send-out. It was almost entirely a math problem, three parts in one hour, and the time was a bit tight. Part 1: Data Center Registration and Health State Management Part 2: Computing the Haversine Distance Part 3: Proximity-Based Routing A few days later I got the usual HR call. They said I did well on the OA, and asked me to briefly go over my…
- OCI Onsite Where the Director Skipped System Design and a Surprise LLD Round, Rejected Without FeedbackOracle · Software Engineer · Onsite · Sep 2026Rejected
JD: a public posting on Oracle's careers site. I mass-applied. Got a recruiter call the very next day, a phone screen a week later, and the onsite another week after that. I interviewed over two days, Wednesday and Thursday. Got told today that I failed, and they won't give any feedback. After talking with them, it felt like there are a lot of openings and they're actively hiring. Phone screen: c…
- Offer After 1.5 Years Out of Work and 600+ Hours of PrepMeta · Software Engineer · Onsite · Sep 2026Offer
I had been out of work for a year and a half. Once I got properly serious about the search, I fired off about 100 applications and basically got silence. A few startups got in touch, but nothing picked up speed. Then someone referred me at Meta and a recruiter asked to set up a call. One choice early on, I believe, defined everything that followed. I turned the startup conversations down and comm…
- One DSA Round, Two Mediums, RejectedAmazon · Frontend Engineer · Technical Screen · Sep 2026Rejected
I applied through a LinkedIn post for SDE I (Frontend). HR contacted me over a call and we discussed the further rounds. Round 1 (PS-DS Round) (1hr) Two DSA medium questions were asked, and I was able to solve both of them within the time limit (~40 mins). The interviewer was helpful and provided hints. Q1. Lowest Common Ancestor of a Binary Search Tree Q2. Ones and Zeroes Explain brute force, th…
- One Hiring Manager Round on Terraform Provisioning, Then a RejectionApple · Software Engineer · Technical Screen · Sep 2026Rejected
The first round was a 55-minute interview. From my own digging before the interview, I figured out that the person interviewing me was the Hiring Manager. The first half was just chatting with them. Every Apple interview I've done so far has been like this. The last 20-some minutes were some high-level design questions. This team does management and maintenance of infra, and the question was very…
- One Virtual Onsite Round with BQ Drill-downs and a Checkout Price Engine OODAmazon · Software Engineer · Onsite · Sep 2026
Three BQ questions: Q1: Can you tell me about a time when you pushed back against a decision that negatively impacted your team? What was the issue? How did it turn out? Q2: Can you tell me about a time when a system was failing intermittently, and the very obvious metrics were all looking good? How did you find the root cause, and how long did it take? Q3: Can you tell me about a problem you had…
- One-Hour AI-Assisted Phone Screen on Camel CardsRippling · Software Engineer · Technical Screen · Sep 2026In progress
The phone screen was one hour, and the question was one of the ones people have already posted about: Camel Cards. I picked the option where the AI helps write the code. A hand has 5 cards, and there are 7 hand types in total, each with its own weight, ordered from strongest to weakest. Part one: first I read the problem and made clear that I would create 1 abstract class HandType and 1 class Han…
- One-Hour Phone Screen with BQ and a CodeSignal Pagination APIOkta · Software Engineer · Technical Screen · Sep 2026
Okta's interviews feel like they vary a lot by team. I read a few other write-ups and they were all pretty different. From the recruiter to the interviewers, the company culture seemed pretty good and everyone was nice. If I get rejected, it just means I wasn't good enough. Anyways, posting this for whoever interviews next. My phone screen was 1 hour: 25 minutes of BQ, 30 minutes of coding, 5 min…
- One-Hour Take-home on Aptora, Rejected the Next DayRo.Co · Applied AI Engineer · Senior+ · Online Assessment · Sep 2026Rejected
Ro.co: there are basically no interview reports for this company on the forum, so I'm sharing a data point. It's based in NYC, a telehealth startup that's been around for several years, and its main competitor is Hims & Hers. I interviewed for a Senior Applied AI Engineer position. My background is traditional SWE, and I applied blindly online. I talked to the recruiter and the HM separately firs…
- One-Hour Technical Screen, Coding Then DesignBNSF · Software Engineer · Technical Screen · Sep 2026
The technical interview was one hour, split into two parts: the first half was writing a program, the second half was design. The problem was provided through a markdown file. It was a requirements doc for a REST API, roughly a price quote (rate inquiry) endpoint for some kind of railway freight, with other dependent services and database access operations involved. I was asked to write a sample…
- Online Assessment Coding Question on Maximizing a Bitwise ANDAmazon · Software Engineer · Online Assessment · Sep 2026
Fresh out of the oven. I found the coding question really confusing. I spent a long time thinking and guessing, and I guessed bitwise AND, which turned out to be wrong. Afterwards I went back over it with Claude, and the approach it gave me was greedy + bit manipulation. Here is the summary: Problem: Given an array Arr, a count K, and a budget C. You can distribute C total "capacity" as increases…
- Online Assessment with a Coordinate-Compression Problem and a Debugging TaskAmazon · Software Engineer · Online Assessment · Sep 2026
One algorithm question, one project question, an AI-coding debug round. Q1: There are infinitely many bags in a row. Each segment [l, r, v] means bags l through r each hold v dollars. Segments do not overlap, and empty bags count as 0. Pick k consecutive bags and find the maximum total amount, mod 1e9+7. Constraints: n
- Online Assessment with a Django Debug RoundAmazon · Software Engineer · Online Assessment · Sep 2026
40 min Coding. It was pretty hard, the approach was greedy + bit manipulation. Feels like Amazon doesn't want to hire anyone, lol 60 min Debug. This one is tricky, so let me roughly go over my approach: The second part is an existing Django project, and you have to debug it based on the failing tests. You are not asked to write Django from scratch. The debug order I'd recommend is: Test → URL → V…
- Online Assessment with a Merge-Strings Conflict Problem and an AI Spring Boot Bug-Fix TaskAmazon · Software Engineer · Online Assessment · Sep 2026
CODE Given two strings (lowercase English letters), primary and secondary, with lengths m and n respectively, try to merge them into one string: Keep the original order within each string A smaller character has higher priority A conflict is defined as the case where a character with low priority is merged before a character with high priority in the final merged string. For example, if the merge…
- Online Assessment with Randomized Probability, Combinatorics and Logic PuzzlesSIG (Susquehanna) · Quantitative Trader · Online Assessment · Sep 2026
The questions are similar to the ones you can find online, but there are some variants. All the specific numbers and shapes are randomly generated. Below is a summary done with AI, in shuffled order. Probability and Bayes' Theorem Factory A produces $a\%$ red parts and $(100-a)\%$ black parts, and Factory B produces $b\%$ red parts and $(100-b)\%$ black parts. A factory is picked at random and $k…
- Onsite with a Cat/Dog Comment-Tree Filter and a Comment-Likelihood ML DesignReddit · Machine Learning Engineer · Onsite · Sep 2026
Coding: seems like a new question? I had AI rewrite it: You are given a dictionary of Reddit comments. Each comment contains: id parent_comment body cat: whether the comment is about cats dog: whether the comment is about dogs Example input: Comments form a tree using parent_comment. There are two user modes: CAT_PERSON: does not want to see dog-related comments DOG_PERSON: does not want to see c…
- Onsite with Grid Design and an Infection Coding RoundOpenAI · Software Engineer · Onsite · Sep 2026In progress
I finished my onsite last week, so I'm sharing my experience before the result comes out. Design: grid. I hadn't looked at any interview reports and was totally clueless. The interviewer was really nice and kept guiding me, and in the end I managed to come up with some stuff, but it's probably not looking good. Project: I talked about a fairly big product that I recently led the launch of. The sc…
- Onsite With Two System Design Rounds and Escalating Follow-upsPinterest · Software Engineer · Onsite · Sep 2026
Round 1 — Behavioral A standard set of leadership questions, one by one: Self-introduction Pick a project where you had the strongest ownership, or learned the most from, and go deep on it How did you measure success for this project? What's the strongest team environment you've been part of? A time you had to make a fast decision and live with the consequences A time you got constructive feedbac…
- Open-Ended System Design That Ran Fifteen Minutes OverGeneral Motors · Software Engineer · Onsite · Sep 2026
The interview was scheduled for 45 minutes. The interviewer gave me a question that was roughly: there is an application that has to call three to five dependency services, and how do you guarantee a specific target for a specific service, for example access latency under 10 seconds and availability above 99%. It didn't feel like a normal design round. It felt like a very open-ended discussion ab…
- Pandas Tasks and Open-Ended Retention PredictionMixpanel · Data Scientist · Technical Screen · Sep 2026In progress
Following up on my previous SQL post: This was the Python screen in the same Mixpanel DS phone-interview round, about 45 minutes. There were four questions: two hands-on pandas tasks and two open-ended questions. Aggregate the users and events dataframes separately. Merge the two dataframes and check for duplicates. An easy one. Open-ended: Which features might predict retention? Explore the data…
- Passed the Coding OA, Failed the Behavioral OAIBM · Software Engineer · Online Assessment · Sep 2026Rejected
Two problems, 75 minutes. The first problem was counting triplets. You're given an array and a number d, and you have to return the number of triplets that satisfy a condition: in each triplet, the difference between the largest number and the smallest number can't be greater than d. My first solution was to sort the whole array and then use two nested for loops to find the smallest and largest n…
