What candidates say the round looked like, summarised in our words, each linked to where it came from. Rippling hasn’t confirmed these details unless the source says “Official page”.
Had this round? Add your report →
Senior loop: recruiter screen, a coding screen, then onsite coding, system design and a project presentation. For the onsite coding round, candidates could choose between an AI-assisted round and a traditional one. This candidate chose to implement the solution themselves without AI.
A 60-minute live round on a business-style problem: build an expense policy engine that takes card transactions, applies configurable rules and returns approved or rejected with a reason. The candidate spent the first 10 minutes on design only, naming the entities and clarifying how rules combine. In the next 25 they had the AI scaffold the data models and boilerplate while writing the core evaluation logic themselves. The last 20 went on edge cases, with AI-generated tests they walked through one by one.
Senior AI coding round. Rippling didn’t limit AI use: the candidate could paste the prompt into Cursor as long as the code went back into HackerRank and local files and chat history were deleted afterwards. They talked through their approach before letting the AI do anything and wrote explicit prompts, so it was clear the AI only sped up the typing. The problem had 3 coding parts and they doubt they could have finished all 3 without AI. Tip: use a fast model, since there’s a lot of simple code to write and a top model is overkill.
San Francisco candidate with an offer. Both the phone screen and the onsite coding rounds allowed any AI tool: ChatGPT, Cursor, Claude or whatever they normally use. Code still had to run in HackerRank, so they copied it across from their own IDE. The questions were not typical LeetCode problems: the algorithms were easy and the work was closer to object-oriented design. They weren’t sure how AI use was graded until the recruiter called with good news.