Point72 Quantitative Researcher interview experience

The setup was tricky: the training data was much smaller than the test data, and there was a large amount of noise—around thirty percent—in the small training set.

That imbalance made the task feel genuinely difficult, because there wasn’t much signal to lean on early and I had to reason carefully about how the model might behave on noisier, unseen data.

I didn’t end up getting an offer. Honestly, the take-home was the hardest part of the experience for me because it didn’t feel like a typical, clean exercise—it was more like a messy real-world scenario designed to test judgment.