United States of America Akuna in the US has recently been sending out its 2026 Online Assessment (OA), especially in the QR, SDE, and MLE tracks; almost every student received an invitation.
This time I took some time to systematize all the questions from the last three sessions (25-27NG).
If you've just gotten your link, or you're not ready to test it yet - don't worry, this article can help you "step in half the hole".
Overall Process and Notes
Akuna's OA doesn't play around, it's fully timed, with a fixed number of questions and strict monitoring.
The types of questions vary greatly from position to position, but the commonalities are time constraints, long questions, and the requirement for precision.
- Release time: Sent out within 1-3 days of online application, tests are usually valid for 1 month;
- systems: Self-developed platform (not HackerRank / CodeSignal), simple UI but no debug mode;
- Type of topic: All algorithmic and logic programming questions, no multiple choice questions;
- language restriction: Python / C++ / Java optional (language will be mandatory depending on the position);
- Test Focus: Algorithmic accuracy, logic clarity, code efficiency.
Never put off doing it until the deadline, the servers jerk around from time to time, especially in the last two days.
QR Summer Intern OA
Duration: 120 minutes / 3 programming questions
Difficulty: Leetcode Easy - Medium
Focus on: mathematical logic + dynamic programming + array traversal skills
The types of questions are relatively constant, examining series, array optimization, DP subproblems, etc.
Sample Questions:
Q1. Minimum Swap
Given an array, determine the minimum number of swaps required to sort it in ascending order.
The idea is to use the index mapping relationship to create a mapping between array elements and their sorted positions, and calculate the number of swaps through cycle detection.
Q2. Array Difference Sum
Given an array, compute the sum of absolute differences between left and right elements for each index.
The classic prefix sum + difference combo problem that focuses on time optimization and O(n) traversal skills.
Q3. DP Subsequence Sum
Find the maximum sum of a subsequence such that no two selected elements are adjacent.
It's actually the revamped House Robber issue, note that you can't jump 2 or more elements in a row.
The core idea is dp[i] = max(dp[i-1], dp[i-2] + nums[i]).
It is always recommended to practice DP and greedy questions in the QR direction, especially the judgment of boundary conditions.
C++ / SDE 2026 NG OA
Duration: 80 minutes / 3 questions
Difficulty: medium to high, extremely tight!
Examination focus: data structure design, system state maintenance, structured thinking
It's not the algorithm that's most likely to drop points on this set, but thetiming. The questions are super long and you have to read each question twice to fully understand it.
Sample Questions:
Q1. String Manipulation / Parsing
Implement a parser that reads input strings with special characters and outputs formatted results according to given rules.
Examining string processing and edge case processing, the questions are verbose but logical.
Q2. Trading Record System
Design a trading record system that maintains both buy and sell orders, including price and quantity.
The system should also handle "cancel" operations and update the book accordingly.
Be sure to pay attention here:
- Two data structures (e.g., two priority queues or maps) need to be maintained for buy and sell orders;
- The total amount is updated with each insert / cancel operation;
- The output requirements are strict, and data format errors will be directly judged wrong.
Q3. Array Simulation
A series of operations are applied to an array - rotation, insertion, and removal. Output the final state after all operations.
This may seem like a simple question, but actually you have to pay attention to input and output format alignment and boundary exceptions.
Familiarity with STL containers is recommended (especially map, priority_queue, deque) efficiency of use.
Quant Research (QR) 26 Summer OA
Duration: Approx. 120 minutes / 3 questions
Difficulty: medium - hard, focus on logical reasoning + mathematical thinking
These questions are obviously closer to "quantitative thinking", not just coding, but embedded in trading scenarios.
Sample Questions:
Q1. Release Schedule Optimization
Given two arrays plannedDate And alternateDate, find the minimum total days to release all updates.
This question examines the greedy scheduling idea, which involves combining sorting and conditional judgments to optimize the order of tasks.
Q2. Array Pair Swap
Given two arrays a And b, find the maximum number of pairs (i,j) Such that a[i] + b[j] ≤ k.
Typical two-pointer + sorting problem where a brute force solution O(n²) would time out.
Q3. Simulation / Expected Value Estimation
Probability-based simulation questions, such as calculating expected returns or risk-adjusted returns from a given distribution, are tested in some editions.
To master basic statistics + use of Python numpy.
MLE / Quant Tech OA (partial version)
Some of the technical posts (such as Machine Learning Engineer or Quant Tech) have questions between QR and SDE.
Usually included:
- An algorithm (DP / Greedy)
- A Matrix Operation (NumPy / Mathematical Modeling)
- One logic or systems question (e.g., caching or order matching)
This type of position is more hybrid profile oriented, especially requiring neat code and clear naming.
Exam Preparation Advice and Timing Strategies
- Time Allocation Recommendations:
You will be able to quickly scan through the 3 questions and judge the difficulty of each question within 5 minutes;
Write clear ideas first, then attack tricky questions. - Code Habits:
The output format strictly adheres to the requirements of the questionnaire, which is automatically determined by the system;
Do not print debug logs in the terminal. - Practice directions:
Recommended for multi-brush Leetcode Medium / Arrays, Sorting, Greedy, DP related topics;
QR students can make up some mathematical modeling and probability distribution type questions.
I got through Akuna OA in one go with Programhelp Pro, and the offer came in seconds that week!
Akuna's OA is not about IQ, it's about "calm + speed + structured thinking".
Its design style is very unique, the questions are not mathematical proofs like investment banks, nor pure algorithmic questions like tech companies, but more like "algorithm + financial logic combination".
If you want a solid pass in the official evaluation, no bugs in your code, and all test cases passed at once.
You can familiarize yourself with the questions in advance by simulating the environment.
What we offer OA Real-time Assist Service Supporting the Akuna platform's untethered online guidance.
Helps you think clearly and code with zero mistakes quickly under high pressure time.
