Just finished a game BCGX OA for the Data Scientist data analysis position was successfully completed within one hour. The overall feeling is: it is not difficult, but it tests more details and engineering habits than pure algorithm questions. BCG will change a few sets of questions every once in a while. The questions are large and the difficulty is okay. The important thing is to write the logic clearly and handle the boundaries well. Let’s break down the ideas and talk about them to give some reference to students who are preparing.
Q1: The cumulative number of visits reaches the target
Given the number of visitors every day, ask on which day "the total reaches or exceeds the target for the first time." Idea: Traverse sequentially and then accumulate. If the target value is not reached, it will be returned. Otherwise, -1 will be returned.
Q2: Problems with electric scooters
You start from 0, walk to the nearest scooter on the right each time, and then ride it for up to 10 units, repeat this process, and ask how far you have ridden in total. Idea: Sort and then divide it into two parts. Starting from the starting point, loop to find the farthest distance to the right of it. Calculate the farthest you can slide this time (not exceeding the end point), then accumulate the results and update the position.
Q3: Battery recycling problem
You have multiple batteries, each with a usage time and charging time. Use them in order, charge them when they are used up, skip them if they are not available, and cycle until they are used for t minutes. Ask how many complete batteries are used in total, and return -1 if all batteries are charging halfway. Idea: Memorize the time of all batteries, then loop through from the initial index, calculate the actual duration, and update the current time and remaining time. If there is no battery after traversing a circle, it will return -1, or if the accumulated time is enough, it will return the number of batteries.
Q4: Restore travel order based on adjacent photos
Each photo gives you two adjacent landmarks, which actually gives you a bunch of "adjacent relationships". This is essentially a chain. Idea: Use the adjacency list to build the graph, then traverse the reconstructed path, walk through all the nodes, and complete the path reconstruction.
If you are not completely sure about OA, don’t take it hard
Now, whether it is BCG X, Amazon, Meta, Google or other major North American manufacturers, the questions are moving in the direction of "detail-intensive + time compression". What really widens the gap is never whether it can do it, but whether it can stabilize the AC in a high-pressure environment.
Many students usually have no problem solving the questions, but when it comes to the official OA, the rhythm is out of control, the boundaries are missed, the format is wrong, and in the end they almost miss it. What is poor is often not ability, but stability on the spot. We do it for the long term OA/VO practical support from major North American manufacturers , familiar with the structure of various question types and common stuck points, AC is not a problem.
