← Back to Directory

Flatiron Health

Mid-sized

Flatiron Health is a healthcare technology company focused on oncology, providing electronic health record software and real-world data solutions to improve cancer care and accelerate research. It is known for its OncoEMR platform and its role in advancing data-driven cancer treatment and clinical trials.

7 interview notes · updated Jul 2026

Flatiron Health·Data Scientist·Technical Phone Screen

Jul 2026
Flatiron Health data scientist interview with a heavy experimentation design question. The whole thing was basically one long case study on metrics frameworks and A/B testing, broken into four parts. Dense, but kind of satisfying if you're into this stuff.
  • You're launching a new recommendation module on a content platform. Define a primary success metric with an exact formula, including randomization unit, numerator, denominator, inclusion/exclusion criteria, and a 28-day attribution window. Also define at least two guardrail metrics with thresholds and explain the trade-offs.
  • What randomization unit and bucketing strategy would you use to avoid contamination across surfaces and sessions? How would you detect and triage Sample Ratio Mismatch and event-loss issues, and what p-value threshold would you use for the SRM test?
  • How would you handle novelty effects and seasonality in this experiment? Describe a ramp schedule, how you'd use pre-period covariates to reduce variance, and propose a minimum test duration rule. Also, how would you control for multiple comparisons across five planned segmentation cuts?
  • Walk through your decision rubric for shipping, holding, or iterating based on metric results. Specifically, what do you do if the primary metric is flat but a guardrail metric regresses?

“This is where I spent most of my energy and still felt like I left things on the table.”

View Post

Flatiron Health·Data Scientist·Technical Phone Screen

Jul 2026
Flatiron Health DS interview was a deep technical grind, basically one massive ML systems question broken into four parts. The scope was way broader than I expected for a single session.
  • You've inherited a churn prediction pipeline (7-day horizon) with messy data and suspected feature drift. Walk through how you'd redesign it end-to-end: temporal slicing to avoid label leakage, a time-series-appropriate cross-validation scheme, and a feature store setup that keeps training and serving consistent.
  • How would you define and select offline vs business metrics for this churn model, and how do you set a classification threshold when the costs of false positives and false negatives are asymmetric?
  • Describe a concrete monitoring plan for data quality and feature drift in this pipeline. How do you distinguish drift in input features from drift in the target label caused by actual product changes?
  • Lay out a rollout plan for this model in production: canary deployment, shadow mode, rollback conditions, retraining cadence, and fairness checks across sensitive user cohorts.

“This is where I spent most of my mental energy.”

View Post

Flatiron Health·Data Scientist·Take-home Assignment

May 2026
Two-part take-home for a Data Scientist role at Flatiron Health, one chunk in R using tidyverse and one in SQL on MySQL 8.0. The problems were more involved than I expected for a screening stage, especially the SQL piece which had a dynamic top-N filter baked in.
  • You're given a Customers CSV with columns for id, signup date, last active date, a churned flag, and pay. Clean the data according to a specific set of rules (drop null ids, fill null pay with 0, infer churned status from pay, compute active days with a floor at zero) and then compute two metrics: average active days across all remaining customers, and average pay among churned customers. Use tidyverse R and round both results to 2 decimals.
  • Using a schema with Teachers, Students, Assignments, and Grades tables, write a single MySQL 8.0 query that computes each student's average score, excludes students with no grades, keeps only the top CEIL(N/2) students by average score (where N is the count of students with at least one grade), breaks ties by student_id ascending, and outputs student_id, name, rounded average score, and a dense rank. The query must remain correct as data grows.

“The rules seemed straightforward at first but rule 3 tripped me up a bit.”

View Post

Flatiron Health·Product Manager·Onsite - Product Sense / Strategy

May 2026
Flatiron Health product sense question framed around a very specific healthcare persona. Short but dense, the kind of question where you either have a mental model for clinical quality workflows or you're winging it.
  • You're a Chief Quality Officer at a hospital using a cancer research tool. What does your dashboard look like?

“This one tripped me up more than I expected.”

View Post

Flatiron Health·Product Manager·Technical Phone Screen

Apr 2026
Interviewed at Flatiron Health, got a product analytics case question about diagnosing a usage drop. Pretty lean on details from my end but the question itself was meaty enough to chew on for a while.
  • One of our tools saw a 20% drop in usage over a two-week period. How would you go about diagnosing what happened?

“I started with the obvious stuff: did anything ship recently, any infra changes, any external events.”

View Post

Flatiron Health·Data Scientist·Technical Phone Screen

Apr 2026
Technical screen at Flatiron Health for a DS role, pretty much entirely stats and experimentation. They went deep on A/B testing fundamentals, which I expected, but the CUPED piece and the mid-experiment analysis together made it a long session.
  • Walk through deriving the per-variant sample size for a two-proportion A/B test from scratch, showing intermediate values like z-scores and the pooled variance term. Baseline conversion is 4.5%, you want 80% power to detect a 10% relative lift at two-sided alpha 0.05.
  • If CUPED reduces variance by 30%, how does that change your required per-variant sample size, and what's the absolute and relative reduction compared to the unadjusted calculation?
  • You have 50,000 users per variant mid-experiment, with observed rates of 4.5% and 5.2%. Run a two-sided z-test for the difference, report the z-statistic, p-value, and a 95% confidence interval for the difference in percentage points. Is it significant at alpha 0.05, and what does it mean practically?

“I've done this derivation before but blanked on the pooled variance step mid-explanation.”

View Post

Flatiron Health·Data Scientist·Technical Phone Screen

Sep 2025
SQL-heavy technical screen for a Data Scientist role at Flatiron Health. Three questions, all window-function-only constraints, progressively nastier. The kind of interview where you finish and genuinely aren't sure if you passed.
  • Given a table of user events (one row per user per day per event), identify each user's first 3-day consecutive activity streak whose end date falls within a specified 7-day window. Return user_id, streak_start_date, and streak_end_date using only window functions and date arithmetic.
  • For each calendar day in a given date range, compute DAU, daily revenue, a 7-day trailing average of DAU, and revenue per active user (with division-by-zero handled as NULL). Use only window functions.
  • Flag users who placed an order on some date t and then had a strict 10-day gap with no events, followed by any event on a later date. Return user_id, the order date, the gap length in days, and the first post-gap event date. Only the earliest qualifying gap per user.

“This one took me a while to even parse.”

View Post