← Back to Directory

Point72 interview notes, page 4 of 4

Large Enterprises

Most of these Point72 write-ups come from Data Scientists (13), then Software Engineers (13) and Machine Learning Engineers (10). The most reported rounds are the technical phone screen (26), the online assessment (5), the hiring manager screen (2) and the behavioral / leadership round (2).

38 interview notes · updated Jul 2026

Point72·Data Scientist·Technical Phone Screen

Jun 2026
Point72 data scientist interview with a pretty gnarly SQL/Python question about date parsing and fiscal quarters. One question, but it had a lot of moving parts and I definitely underestimated how much edge case handling they expected.
  • Given an 8-digit integer date key in YYYYMMDD format, write both a SQL expression and a Python function that convert it to a quarter label like 'YYYY-Qn'. The solution must validate the date (returning NULL or raising an error for impossible dates like month 13 or Feb 30), support an optional fiscal start month parameter, and handle string inputs with whitespace or leading zeros correctly.

“I thought I had this in the bag because I've written quarter logic before, but the fiscal year offset tripped me up badly.”

View Post

Point72·Data Scientist·Technical Phone Screen

Jun 2026
Point72 Data Scientist interview with a geometry-based coding problem framed around a graphics engine use case. Pretty applied for a DS role, felt more like a software engineering screen than anything stats-related.
  • Given an array of circle pairs, each defined by center coordinates and radius for both circles, determine for every pair whether the circles intersect, touch, or are completely separate.

“The hint is basically the whole solution once you see it: compute the distance between centers and compare it against the sum and absolute difference of the radii.”

View Post

Point72·Software Engineer·Online Assessment (OA)

Jun 2026
Point72 gave me a PySpark coding problem as part of their data engineering interview. It was a self-contained OA with a skeleton codebase and unit tests to pass, which I actually prefer over live coding with someone watching you.
  • Implement a PySpark job class for bank data mining. The class needs four methods: one to initialize a local SparkSession, one to join accounts and transactions filtering out rows where the transfer amount exceeds the source account balance or the destination account doesn't exist, one to return the count of distinct source account numbers, and one to return the top 10 source accounts by transaction count as a dictionary.

“The filtering logic in the second method is where I spent most of my time.”

View Post

Point72·Machine Learning Engineer·Hiring Manager Screen

Jun 2026
Screening call for an ML Engineer role at Point72, where they're spinning up a new team around agentic AI. The interviewer was clearly filtering hard for people who've done real pretraining work, not just fine-tuning.
  • Have you trained a Transformer or LLM fully from scratch? Walk through the whole thing: data pipeline, architecture decisions, distributed training setup, optimizer and scheduler choices, evaluation, and any debugging you had to do.
  • If your background is mostly in parameter-efficient fine-tuning methods, how do you articulate what you've actually done versus what you haven't? Where exactly is the line for you?
  • If you were hired despite not having full pretraining experience, what would your ramp-up plan look like?

“This was the whole conversation basically.”

View Post

Point72·Software Engineer·Online Assessment (OA)

Jun 2026
Point72 data engineer assessment, two independent coding tasks sent as an OA. One SQL, one PySpark. Nothing behavioral, just code.
  • Write a SQL query that, for each candidate, shows the top three distinct vote totals by state, including state names and counts formatted into place columns (1st, 2nd, 3rd). States tied at the same vote count should be grouped together and sorted alphabetically.
  • Implement several methods of a PySpark job class: initialize a Spark session, filter transactions to only valid ones based on account existence and balance rules, count distinct source accounts, and return the top 10 source accounts by transaction count as a dictionary.

“This one took me longer than it should have.”

View Post

Point72·Data Scientist·Technical Phone Screen

Jul 2026
Point72 Data Scientist interview with a SQL-heavy technical screen. One meaty query question that covered window functions, nulls, and tie-breaking logic all at once. Felt like a real test of whether you actually know SQL or just know the basics.
  • Given a Departments table and a Students table (with nullable dept_id, nullable GPA, and an enrollment date), write a single SQL query that returns every department including ones with no students, along with their top student by GPA. Ties should break by earliest enrollment date, then by smallest student ID. If a department has no students with a non-null GPA, return that department row with nulls for all student fields.

“This one took me a minute to untangle.”

View Post

Point72·Data Scientist·Technical Phone Screen

Jul 2026
Point72 Data Scientist interview with a SQL/Python question around date parsing for financial reporting. Pretty focused technical screen, nothing too wild, but the specifics tripped me up a bit.
  • You have a transactions table where dates are stored as 8-digit integers in YYYYMMDD format. Write SQL and/or Python to convert each value into a quarterly label like '2023_Q1'.

“I knew the general idea immediately but fumbled the month-to-quarter mapping under pressure.”

View Post

Point72·Data Scientist·Technical Phone Screen

Jul 2026
Point72 data scientist interview with a geometry-heavy coding problem that felt more like a competitive programming round than anything I expected from a finance firm. The problem was well-scoped but the edge case handling is where they really pushed.
  • Given an array of circle pairs where each circle is defined by (x, y, r), classify each pair as IDENTICAL, CONCENTRIC, TOUCHING_EXTERNALLY, TOUCHING_INTERNALLY, INTERSECTING, or DISJOINT. Coordinates and radii are integers up to 1e9, and N can be up to 2e5. You must avoid floating-point errors, handle degenerate cases like r=0, and explain your time and space complexity.

“My first instinct was to just compute Euclidean distance and compare, which the interviewer immediately flagged.”

View Post