Scale AI · OA → Phone screen → Onsite
In processScale AI New Grad SWE and MLE Question Bank
Scale AI new grad SWE and MLE question bank, 2019–2026: 30 questions and projects from the OA, phone screen, onsite, debugging and design rounds.
At a glance
- What this is: a question bank, not a set of candidate stories. It groups the 30 Scale AI new grad questions on FastPrep by the round they came from, with every date each was seen.
- Process seen most recently: a long, multi-part phone-screen build (task processor, card game), then an onsite with a debugging round and a backend practical.
- Seen most often: Task Processor: Dependencies and Deadlines (5 sightings, latest Aug 2026) and Minimum Distance Matching in a Forest (5, latest Feb 2024).
- MLE differences: data loading, model evaluation, training-data debugging and statistics tasks, such as Fit Beta Parameters by Maximum Likelihood.
Coverage: Scale AI new grad software engineer and machine learning engineer questions, 2019–2026: OA, phone screen, onsite coding, debugging, backend and ML practicals, system design and low-level design. For the candidates' own accounts, see the firsthand reports under Keep practicing.
How the rounds fit together
The question bank is broad, but the recent SWE process puts a lot of weight on practical implementation.
- The current phone screen is usually one long build. The task grows in parts, so clean code matters as much as getting the first part right. The task processor and four-player card game are the two strongest recent examples.
- The current onsite has two technical rounds. Day 1 is a one-hour debugging round and a one-hour backend practical. The contractor assignment repo has appeared more than once, and the same setup has stayed very similar across years.
- The OA can still be algorithm-heavy. Arrays, dynamic programming, intervals, trees, graphs, and simulation all show up.
- MLE has a different problem mix. Expect data loading, model evaluation, debugging training data, tree or forest distance work, statistics, and small ML pipelines.
- Older questions are still useful, but newer ones should come first. The best order is the 2026 phone-screen and practical work, then the repeated graph and simulation questions, then the older OA set.
What to practise
- Task Processor: Dependencies and Deadlines — Seen: Aug 14, 2026; Jun 14, 2026; Apr 30, 2026; Mar 27, 2026; Mar 12, 2026
- Four-Player Trick-Taking Card Game Simulation — Seen: Aug 31, 2026; Dec 1, 2023
- Repair Contractor Project Assignment — Seen: Apr 21, 2026; Oct 31, 2024
- Build a CSV-to-JSON LLM Categorization Service — Seen: Aug 1, 2026; May 9, 2025
- Format and Validate a Model Response — Seen: Apr 30, 2026
The first two match the long, multi-part phone-screen style. The next two cover the current debugging and backend practical format. The last one is a good short exercise for parsing, validation, and careful output handling.
Online assessment: 9 problems
| Problem | Level | All seen dates | Reports |
|---|---|---|---|
| Free Windows Between Busy Intervals | Medium | Apr 13, 2026 | 1 |
| Count Stable Server Subsegments | Medium | Nov 13, 2024 | 1 |
| Distinct OR Values of Increasing Subsequences | Hard | Nov 13, 2024 | 1 |
| Running Through Walls with Limited Energy | Hard | Sep 16, 2024 | 1 |
| Complete a Small CNN Oriented-Box Regression Pipeline | ML project | Dec 21, 2023; Sep 23, 2022; Jan 8, 2021 | 3 |
| Maximum Distinct Values After Cross-Array Swaps | Easy | Jun 2, 2023 | 1 |
| Minimize Adjacent Squared-Difference Cost with One Insertion | Medium | Jun 2, 2023 | 1 |
| Minimum Falling Path Sum | Medium | Jun 2, 2023 | 1 |
| Platform Referral Rewards | Medium | Nov 1, 2022; Oct 31, 2022 | 2 |
Two OA prompts are especially easy to miss:
- Count Stable Server Subsegments: count contiguous segments with at least three values where both endpoints equal the sum of everything between them.
- Distinct OR Values of Increasing Subsequences: return every different bitwise-OR value that can be made from a strictly increasing subsequence. Include
0for the empty subsequence and sort the result. The reported value bound was below1024.
Phone screen: 8 problems
| Problem | Level | All seen dates | Reports |
|---|---|---|---|
| Four-Player Trick-Taking Card Game Simulation | Hard | Aug 31, 2026; Dec 1, 2023 | 2 |
| Task Processor: Dependencies and Deadlines | Medium | Aug 14, 2026; Jun 14, 2026; Apr 30, 2026; Mar 27, 2026; Mar 12, 2026 | 5 |
| Free Time for Every Meeting Room | Medium | Apr 13, 2026 | 1 |
| Order Queries by Shortest Processing Time | Medium | Apr 13, 2026 | 1 |
| Overlap Duration of Two Time Intervals | Easy | Jun 15, 2025 | 1 |
| Extract the Prefix of a Filename | Easy | Feb 21, 2024 | 1 |
| Three-Stage Labeling Workforce Simulation | Hard | Sep 19, 2021; Nov 20, 2020 | 2 |
| Crawl Reachable URLs with Concurrent Workers | Medium | Sep 30, 2019 | 1 |
The important pattern is not one data structure. It is handling a task that keeps growing. A clean model, small helper functions, and quick tests make the follow-ups much easier.
Seen in both phone and onsite rounds: 2 problems
| Problem | Level | All seen dates | Reports |
|---|---|---|---|
| Best Poker Category with a Joker | Medium | Mar 4, 2025; Oct 31, 2024 | 2 |
| Minimum Distance Matching in a Forest | Hard | Feb 25, 2024; Feb 21, 2024; Jan 18, 2024; Jan 6, 2024; Dec 21, 2023 | 5 |
The forest problem is one of the most repeated items in the whole bank. The poker problem tests careful rule handling and edge cases more than a hidden trick.
Onsite coding and ML: 5 problems
| Problem | Level | All seen dates | Reports |
|---|---|---|---|
| Stateful Neuron Matrix Update | Medium | Apr 30, 2026; Apr 19, 2026 | 2 |
| Format and Validate a Model Response | Medium | Apr 30, 2026 | 1 |
| Rotate an Array to the Right | Easy | Oct 31, 2024 | 1 |
| Shortest Distance Between Two Nodes | Medium | Jan 17, 2024; Dec 19, 2023 | 2 |
| Fit Beta Parameters by Maximum Likelihood | Hard | Sep 9, 2022 | 1 |
These range from basic array work to stateful matrix updates, graph distance, response validation, and numerical estimation. For MLE roles, being able to explain the math and test the implementation matters just as much as finishing the code.
Debugging and backend practicals: 4 projects
| Project | Format | All seen dates | Reports |
|---|---|---|---|
| Build a CSV-to-JSON LLM Categorization Service | 3 tasks | Aug 1, 2026; May 9, 2025 | 2 |
| Repair Contractor Project Assignment | 1 debugging task | Apr 21, 2026; Oct 31, 2024 | 2 |
| Debug Missing Model Output | 1 debugging task | Apr 13, 2026 | 1 |
| Training Data and Model Error Debugging | 2 tasks | Mar 5, 2024; Jan 18, 2024 | 2 |
The contractor project is the main debugging practice for the current loop. It uses an existing codebase with contractors, courses, projects, and CSV data. The bugs sit in real program behavior: capacity checks, priority order, course filtering, and state that is not reset correctly.
The backend project is different. It asks you to build working behavior across several parts, not just find planted bugs. Practice reading the existing structure first, then make the smallest change that gets one requirement working and tested.
System design and low-level design: 2 problems
| Problem | Format | All seen dates | Reports |
|---|---|---|---|
| Design an Insurance Claims RAG Agent | 4 guided steps | Apr 13, 2026 | 1 |
| Design and Implement a Task and Worker Assignment Model | 5 guided steps | Sep 9, 2022 | 1 |
These are more role-dependent than the coding and debugging rounds. The first covers an end-to-end retrieval and generation system for claims work. The second focuses on objects, assignment rules, state changes, and a clean model for tasks and workers.
Keep practicing
- Firsthand Scale AI new grad reports: two-day onsite (2026), task scheduler phone screen, card game interview, virtual onsite, debugging interview and 2026 OA.
- Scale AI coding question bank and project coding bank (sign-in needed).
- Scale AI interview guide and hiring insights.
- Experience feed: more interview reports from other companies and roles.
Sources: Anywhere useful interview info can be found.
Comments
0No comments yet — be the first to share what you know.