Four rounds of interviews, 1. Interview with regional partner regarding a tech case interview 2. Live Coding interview. Focus on basic data structure and algo qs 3. Another round of tech case interview with principal FDE 4. Interview with MDP, including cultural fit question and a short tech case interview.

BCG Interview Experience (Data Science / Analytics) Round 1: CodeSignal Assessment End-to-end classification case study. Round 2: HR Screening Standard introductory conversation. Questions included: "Tell me about yourself." General background and experience discussion. Round 3: Technical + Case Interview Coding Questions Write code to calculate the number of missing values in a specific DataFrame column. Implement a function to calculate: Average flight duration Average ticket price Number of flights between a given departure and destination Behavioral / Experience Questions Do you have experience moving a solution into production? What would you consider one of your strengths and one of your weaknesses from a career/technical skills perspective? Case Study The case involved a construction company that wanted to prioritize materials that were likely to be delayed. Available data included: Intended delivery date Actual delivery date Cost Weight per unit Material name Material description Several additional features that I don't fully recall I approached the problem as a classification task and discussed the typical machine learning workflow. Some of the follow-up questions included: How would you handle features with high cardinality? Why would you remove highly correlated variables? If correlated variables contain useful information, how could you preserve that information? I suggested feature engineering (e.g., creating ratio-based features). We also discussed: Model evaluation metrics Feature selection considerations General modeling choices Toward the end, the interviewer challenged my framing of the problem and suggested it might be better approached as a forecasting problem rather than a classification problem. They then asked how my approach would change if the problem were framed as forecasting, specifically: How train/test splitting would differ Whether I would account for time dependencies Why random shuffling would no longer be appropriate Overall Thoughts The interview process felt somewhat unstructured to me, and I struggled to communicate my thought process in a clear, step-by-step manner. Most of that was probably on me rather than the interview itself. One thing that surprised me was how technical the interview was. Many online resources emphasize the business and consulting aspects of BCG case interviews, but my experience was much more focused on data science and machine learning concepts. Good luck!