Interview guides · GSK

GSK Data Scientist Interview Questions & Guide 2026

Practice explaining analysis before a GSK data scientist interview: the recommendation, the uncertainty, the follow-ups.

The 60-second brief

A Data Scientist at GSK operates at the intersection of cutting-edge computational science and global healthcare. In this role, you are not simply building predictive models; you are directly contributing to a pipeline that discovers, develops, and delivers medicines and vaccines to millions of patients worldwide. Your work directly impacts how GSK optimizes its clinical trials, understands disease biology, and refines its commercial strategies to ensure life-saving treatments reach the right healthcare providers and patients efficiently.

Scope your preparation by the data you would actually touch, because the title will not tell you. A seat that lives in event logs and weekly readouts rewards fluency in aggregation and metric definitions; a seat that owns a model in production rewards fluency in train/serve skew, retraining cadence and drift monitoring. The fastest way to find out which one you are interviewing for is to ask what the team shipped last quarter and what it gets paged about.

Focus areas

Reconstruct drug exposure intervals from dispensing records
Minimise protected health information in every extract
Build member-month denominators from overlapping coverage spans

Overview

At GSK, data science is divided into several highly impactful domain spaces, ranging from research and development (R&D) to global commercial operations. Depending on your aligned business unit, you might leverage advanced machine learning to analyze genomic data, design natural language processing (NLP) pipelines to extract insights from clinical literature, or build sophisticated segmentation models for healthcare professionals (HCPs). This strategic variety means your technical solutions will directly influence multi-million dollar resource allocation decisions and accelerate drug discovery timelines.

The scale of data and the complexity of the problems at GSK require a unique blend of technical mastery, scientific curiosity, and ethical responsibility. As a Data Scientist here, you will collaborate with cross-functional panels of clinicians, biostatisticians, epidemiologists, and commercial leaders. Succeeding in this role means translating complex, ambiguous biological or market data into actionable, structured insights that align with GSK's core mission to get ahead of disease together.

The GSK interview process

GSK candidates report 4 rounds · ≈ 3-5 weeks. The stages below are what candidates describe, not a published process.

Initial Screening

Most candidates lose this call inside the first two minutes, during the walkthrough of their own background. The account runs chronologically, sits at the level of tools and titles, and never arrives at a decision anyone could have disagreed with. Anchor on a problem instead of a timeline: what the team could not answer, what you did about it, what happened next. Ninety seconds is enough, and stopping on time leaves room for the half of the call that belongs to you. What you ask about how work gets prioritised signals your level more reliably than the walkthrough does.

What to demonstrate

Whether your background summary has a shape (problem, decision, consequence) or is a chronological list of tools and employers
Whether you can account for gaps, short stints and the reason you are looking, unprompted and without hedging
The substance of the questions you ask back, which an experienced screener reads as a level signal

How to prepare

Time your opening walkthrough against a clock. If it runs past two minutes, compress the earliest role into a single clause and spend the recovered time on the most recent one
Write one honest sentence for every gap or short stint visible on your resume and offer it before being asked about it
Prepare questions about how work arrives and gets prioritised: who writes the request, how often priorities change, and what happens to an analysis after it is delivered

Online Assessments

This round decides whether someone can hand you a schema and a question and trust the number that comes back. Correctness under a clock is the bar, not clever syntax. The habit that separates strong from weak answers is checking the grain: after every join, know how many rows you expect and whether the count moved. Most wrong answers in this format are not wrong logic, they are a fan-out from a key that turned out not to be unique, or a filter applied before an aggregate when it belonged after. Say what you expect before you run it.

What to demonstrate

Whether your row counts survive each join, and whether you notice on your own when they do not
Deliberate handling of rows that fail to match, including whether the question needs an inner join or a left join with the non-matches kept and counted
Whether NULLs are treated on purpose, given that a NULL compares equal to nothing and that COUNT of a column skips it
Reaching a defensible answer inside the window instead of a refined one after it

How to prepare

Take a two-table schema, write a join that fans out on purpose, then fix it by collapsing the many-side to one row per key before joining. Repeat until the fix is reflex rather than recall.
Write a funnel as one query and print the distinct user count at each stage, then confirm each stage is a subset of the one above it rather than assuming it
Do a few timed runs in a plain text box with no autocomplete and no formatter, since assessment editors often have neither

Assessment Center/Panel Interview

A day of back-to-back interviews samples your floor, not your ceiling. Four hours in, the habits that carry a good answer are the first to go: restating the question before solving it, asking what the data would have to look like, checking a number before quoting it. What the day decides is whether the tired version of you is still someone to leave alone with an ambiguous problem. The round that sinks a candidate is usually not the hardest one. It is the one immediately after the round that went badly.

What to demonstrate

Whether the late rounds get the same clarifying questions as the first one, or whether you start answering immediately to save effort
Whether a weak answer stays in the room it happened in, instead of following you into the next conversation as apology or distraction
Whether the quality of your questions holds up, since fatigue removes curiosity about the problem before it removes knowledge of the method

How to prepare

Rehearse the length, not just the content: book four mock interviews of different types in one afternoon with short gaps, because the one you need to observe is the fourth
Put the two or three questions you ask at the start of any problem on a card in front of you, so that under fatigue it is a habit you run rather than a decision you make
Decide in advance what the gap between rooms is for: water, one line of notes on anything you promised to follow up, and an explicit close on the round that just ended so it does not travel
Prepare a different closing question for each interviewer, so the end of a long day does not produce the same one four times

Final Discussions

Where a loop ends with a senior leader, that conversation is rarely another skills test. The technical signal already exists by then, so the questions tend to open up: what you would look at first, where a metric you have heard about could mislead, what you would push back on. The decision being made is scope, which in practice means level and how much you would be trusted to own unsupervised. Treating it as a formality is the usual mistake. An open question late in the day is still being scored, and a vague answer reads as someone who has not run anything themselves.

What to demonstrate

Whether your view of the business has anything specific behind it, given that you are working only from what is public and are expected to say so
Whether the scope of work you describe owning matches the scope of the role, instead of sitting a level below it
Whether you can disagree with something concrete and stay useful about it, rather than agreeing with everything said in the room
Whether your questions are ones only this person could answer, as opposed to ones the recruiter already covered

How to prepare

Build one view you could defend for two minutes using only public information: what the funnel probably looks like, which metric likely drives decisions, and where that metric could mislead. Being wrong for a stated reason survives this round; having no view does not
Write down the largest piece of work you have owned from question to decision, who else touched it, and what you decided alone, then check that it reads at the level you are interviewing for
Prepare one thing you would want changed if you joined and phrase it as a question rather than a verdict, so it opens a conversation instead of closing one

Where candidates lose points

Rates built on member counts rather than exposure

Members join and leave mid-period, so dividing events by distinct members mixes a person covered for 30 days with one covered for 365. New joiners also have artificially low observed utilisation because their claims have not arrived yet and because care takes time to initiate. Denominators must be member-months or member-years, and comparative quality measures usually need a continuous-enrolment requirement with an explicit allowable gap, stated in days.

Immortal time in adherence, treatment, and enrolment definitions

Classifying members as adherent, treated, or programme-enrolled requires them to survive and stay covered long enough to accumulate the defining events. That guaranteed event-free interval is assigned to the exposed group, so the exposure looks protective for reasons that have nothing to do with the treatment. Adherence studies are the classic case: measuring 12-month proportion of days covered and then comparing mortality builds survival into the exposure definition. Use time-varying exposure or a landmark analysis with the classification window excluded from follow-up.

Generalising beyond the population the sample actually supports

State the frame the sample was drawn from and where it diverges from the population you want to talk about: time window, platform, geography, opt-in. If a group is excluded from the frame, either weight to known margins or narrow the claim rather than quietly extending it.

Accepting a metric definition without asking about the denominator

Pin down the denominator, the eligibility filter and the time window before computing anything: conversion rate per session, per user, per eligible user and per new user are four different numbers with different behaviour. Restate the definition in one sentence and get agreement before you analyse.

Questions & practice

How would you approach a resource allocation problem using predictive modeling to ensure optimal distribution of medical resources?

Medium · statistics and probability

Approach

Translate the result into the decision it informs, in one plain sentence.
Quantify uncertainty explicitly rather than reporting a point estimate alone.
Say what the estimate is of, and over what population it generalises.

Follow-up

How would you explain this result to someone who does not know statistics?
Which assumption here is most likely to be violated in practice?

Walk me through a time you deployed a machine learning model that directly solved a complex business problem. What metrics did you use to evaluate success?

Medium · machine learning and modelling

Approach

Frame the prediction: the label, the moment of prediction, and the action it triggers.
Set a baseline first, so any model has something honest to beat.
Check what information would not exist at prediction time, and exclude it.

Follow-up

Where could label leakage enter this setup?
What would you monitor after launch to know the model is still valid?

Explain how you would design a clustering algorithm for HCP Segmentation to optimize marketing outreach.

Medium · machine learning and modelling

Approach

Pick an evaluation metric that matches the cost of each error type, not a default.
Set a baseline first, so any model has something honest to beat.
Frame the prediction: the label, the moment of prediction, and the action it triggers.

Follow-up

How would you choose the decision threshold, and who owns that choice?
What would you monitor after launch to know the model is still valid?

What are the key architectural differences between traditional recurrent neural networks and modern transformer-based models in NLP?

Medium · machine learning and modelling

Approach

Say how the offline result would be validated online before it is trusted.
Check what information would not exist at prediction time, and exclude it.
Pick an evaluation metric that matches the cost of each error type, not a default.

Follow-up

What would you monitor after launch to know the model is still valid?
How would you choose the decision threshold, and who owns that choice?

Collapse claim versions before summing allowed amounts by member

Easy · claim versioning · groupby · deduplication

You are given medical_claim_line as a pandas DataFrame: claim_line_id, claim_id, claim_version, frequency_code (1 original, 7 replacement, 8 void), member_id, service_start_date, procedure_code, allowed_amount, claim_status. Return total allowed_amount per member for service dates in one stated calendar month, counting each claim once at its surviving version and excluding any claim whose surviving version is a void. Multi-line claims must keep every line of that surviving version. Output columns: member_id, allowed_amount, sorted descending. Do not deduplicate on claim_id alone.

Approach

Separate the two grains out loud before writing code: versions live at the claim level, dollars live at the line level. Every bug in this exercise comes from mixing them.
Compute the surviving version per claim with a groupby transform of max over claim_version, then keep the lines whose claim_version equals that value. A transform keeps the frame at line grain, which a sort plus head(1) does not.
Drop whole claims whose surviving version carries frequency_code 8. A void is not a zero-dollar line, it retracts the claim, so filtering rows rather than claims leaves the prior version's lines behind.
Filter to claim_status 'paid' and state the assumption: denied, pended and reversed lines carry an allowed_amount that never became a liability, so including them overstates cost.
Restrict service_start_date to the month boundaries, then groupby member_id and sum. Sort descending and return.

Worked solution (20 min)

Build a max-version lookup per claim_id with groupby('claim_id')['claim_version'].transform('max') and keep rows where claim_version equals it.
From the surviving rows, collect the claim_ids whose frequency_code is 8 and drop those claim_ids entirely.
Filter claim_status to 'paid' and service_start_date into the month window.
Group by member_id, sum allowed_amount, sort descending, reset the index.

Expected result: A two-column frame member_id, allowed_amount. On a claim with versions 1, 7, 7 only the third version's lines contribute; a claim whose latest version is frequency_code 8 appears nowhere. Total allowed is strictly below a naive sum over all rows whenever any claim in the sample was adjusted, and equal to it when none was.

Verify it yourself

The distinct claim_id count in your filtered frame equals the count of distinct claim_ids whose surviving version is not a void.
The line count you kept for each surviving claim equals that claim's line count at its max version, not 1.
A naive sum over every row is greater than or equal to your total, with equality only when no claim has claim_version above 1.

Follow-up

A claim has versions 1, 7, 7 where the second replacement has fewer lines than the first. What does your code return and is that correct?
How would you handle a claim whose lines span two calendar months at the boundary of the reporting window?
The same logic has to run over 400 million lines in a warehouse. What changes, and what stays the same?

Compute allowed PMPM by incurred month without span fan-out

Hard · fan-out · many-to-many · runout

Report allowed PMPM by incurred month and product_type for 2025 from medical_claim_line (member_id, claim_id, claim_version, frequency_code, service_start_date, allowed_amount, claim_status), pharmacy_claim (member_id, fill_date, allowed_amount, reversal_flag, reversed_fill_id), member_enrollment (span_id, member_id, product_type, effective_date, termination_date) and member_risk_score (member_id, score_year, prospective_risk_score), which holds at most one row per member per score year. The numerator is allowed_amount on surviving claim versions plus unreversed fills, attributed to the product_type in force on the service or fill date. The denominator is member-months. Also return a risk-adjusted PMPM for each cell: allowed PMPM divided by the member-month-weighted mean prospective_risk_score for score_year 2025 over the members making up that cell's denominator. A colleague joined claim lines to member_enrollment on member_id alone and the annual total came back 1.4 times the finance figure. Diagnose that, then produce the correct monthly series.

Approach

Name the defect precisely. member_enrollment is one row per span, so joining on member_id alone is one-to-many and duplicates every claim line once per span that member ever held. The inflation factor is the spend-weighted mean span count per member, which is why 1.4 is plausible and why it is not a constant you can divide back out.
Make the join functional: member_id AND service date between effective_date and COALESCE(termination_date, DATE '9999-12-31'). That yields at most one span per line unless spans genuinely overlap, in which case pick a documented precedence rule, most recently effective wins being the usual one, and prove uniqueness with a duplicate check rather than asserting it.
Aggregate numerator and denominator separately and join them on (month, product_type). Computing both sides across one join is the second route to inflation, because the member-month denominator fans out over claim lines just as easily as the dollars fan out over spans.
Clean each numerator source on its own terms: surviving claim versions with voids removed and claim_status 'paid' on the medical side, reversal rows and their originals removed on the pharmacy side.
Apply completion before anyone reads a trend. State the paid-through date and either withhold the three most recent incurred months or apply completion factors, since the lag differs by service type and the incomplete tail reads as a cost improvement and a utilisation drop at once.
Risk-adjust last, and weight the score the way the denominator is weighted. The divisor for a cell is SUM(member_months * prospective_risk_score) / SUM(member_months) over that cell's members, not a plain AVG over distinct members: a member covered two months must not count the same as one covered twelve. member_risk_score is member-level, so join it after the member-month grain is fixed, or it becomes a second fan-out.
State the rule for members with no 2025 member_risk_score row: keep them in member_months, leave them out of the weighted mean, and publish the scored share of member-months beside the adjusted series. A thin score table moves the adjusted number while the raw one sits still, and without that share nobody can tell the two apart.

Follow-up

A member holds two overlapping spans under different product_types on the service date. Which one gets the dollar, and how do you keep the member-month denominator consistent with that choice?
Write the check that proves the span join is one-to-one rather than assuming it.
Finance reports on paid_date and you reported on service_start_date. Reconcile one month between the two bases.

Deduplicate claim versions before totalling allowed amounts

Easy · dedup · window functions · claim versions

medical_claim_line holds one row per service line per claim version: claim_line_id, claim_id, claim_version, frequency_code (1 original, 7 replacement, 8 void), member_id, service_start_date, procedure_code, allowed_amount, claim_status and adjudicated_at. Adjusted claims appear more than once, so summing every row double counts them. Return total allowed_amount by procedure_code for service_start_date in Q1 2025, counting only lines with claim_status 'paid' on the surviving version of each claim, which is the highest claim_version per claim_id, and dropping the claim entirely when that surviving version carries frequency_code 8. Resolve survivorship with a window function, not a self-join.

Approach

Resolve survivorship at claim_id, not claim_line_id. Carry MAX(claim_version) OVER (PARTITION BY claim_id) alongside every line and keep lines where claim_version equals it, because a replacement version can contain a different number of lines than the original.
Read the frequency_code of the surviving version only. A frequency_code 8 removes the claim outright; it does not revert payment to the prior version, so a void must delete the claim rather than promote version n-1.
Apply claim_status and the date filter after survivorship is resolved. Filtering paid lines first can strip the surviving version and silently elect a superseded one.
Group by procedure_code and sum allowed_amount, then state the paid-through date on the output because Q1 amounts keep moving until adjudication runout completes.
Sanity check the shrinkage: report how many claim_ids and how many dollars the dedup removed, so the reviewer can see the step did something.

Worked solution (20 min)

Build a CTE that adds surviving_version = MAX(claim_version) OVER (PARTITION BY claim_id) and surviving_freq = the frequency_code of the row at that version, to every line.
Filter to claim_version = surviving_version so all lines of the surviving version survive together.
Drop claims whose surviving_freq is 8, then apply claim_status = 'paid' and service_start_date between 2025-01-01 and 2025-03-31.
Group by procedure_code, sum allowed_amount, and emit the count of distinct claim_ids contributing.
Run the naive SUM over all rows beside it and report the difference as the double count that was removed.

Expected result: One row per procedure_code with summed allowed_amount over surviving, non-voided, paid Q1 2025 lines. A claim with an original and one replacement contributes only the replacement's lines; a claim whose surviving version is a void contributes nothing.

Verify it yourself

Deduped total is strictly below the naive paid-and-Q1 SUM taken across all versions whenever at least one claim carries paid Q1 lines on more than one version. If the two totals match on such a claim, the dedup did nothing.
For one claim_id known to have three versions, the line count it contributes equals the number of its paid Q1 2025 lines at its highest claim_version, not 1.
Let B be the set of claim_ids whose surviving version has at least one paid line with a Q1 2025 service_start_date. COUNT(DISTINCT claim_id) in the output equals the size of B minus the number of claims in B whose surviving version carries frequency_code 8. Comparing instead against COUNT(DISTINCT claim_id) over the whole raw table is not a void test: that gap also absorbs every claim dropped by claim_status or by service date.

Follow-up

Two rows share a claim_id and claim_version but differ on adjudicated_at. What do you do, and what does that imply about the extract?
How does the query change if the request is allowed_amount by paid_date rather than service_start_date, and which basis does a finance reconciliation want?
A void arrives for a claim already published in a closed month. How do you restate without rewriting history in the warehouse?

Describe a technical project you contributed to where you had to learn a new tool or methodology very quickly. What was your approach?

Medium

Approach

Say what you would check first and why it is the highest-information step.
Clarify what is being asked and what a complete answer would contain.
Work from the decision backwards to the evidence you would need.

Follow-up

How would you know your answer was wrong?
What assumption would you test first?

Decompose risk-adjusted cost per member into owned leaf metrics

Hard · metric tree · price volume mix · coding intensity · ownership

Build a metric tree, at most four levels deep, from risk-adjusted allowed PMPM down to leaves an owner can move. Data: medical_claim_line (allowed_amount, units, procedure_code, place_of_service_code, service_start_date, claim_version, frequency_code), pharmacy_claim (allowed_amount, days_supply, is_generic, formulary_tier, reversal_flag), member_enrollment (coverage spans, prospective_risk_score). Give the arithmetic identity each split rests on, the leading indicator at each leaf, and identify the one node where the arithmetic is exact but the attribution is not.

Approach

State the base identity before drawing anything. Allowed PMPM equals total allowed over member-months, and it decomposes exactly by service category c as the sum over c of (units_c / member-months) times (allowed_c / units_c). Utilization per member-month and allowed per unit, summed across categories, so category mix is inside the sum rather than a separate hand-wave.
Give the change decomposition too, because the tree will be read on deltas: the change in PMPM equals the sum over categories of (change in utilization times base price) plus (base utilization times change in price) plus (change in utilization times change in price). The third term is the interaction, it is not noise, and an analysis that drops it will not reconcile.
Build the medical branch to leaves with owners: inpatient admits per 1,000 member-months and allowed per admit (with DRG case mix as the sub-split), ambulatory visits per 1,000 and allowed per visit split by place_of_service_code, and emergency visits per 1,000 and allowed per visit. Prepare the data with version dedup to the surviving claim_version before a single dollar is summed, since frequency_code 7 replaces and 8 voids.
Build the pharmacy branch on 30-day equivalents: scripts_30 equals the sum of days_supply over 30 on unreversed fills, so pharmacy PMPM equals (scripts_30 per member-month) times (allowed per 30-day script). Split the price term exactly as g times the generic price plus (1 minus g) times the brand price, where g is the generic dispensing rate, which makes generic substitution and unit price separable leaves with different owners.
Attach one leading indicator per leaf: prior-authorization approvals and elective surgical bookings ahead of inpatient admits, contract effective dates ahead of allowed per unit, site-of-service steering ahead of the ambulatory place_of_service mix, new-to-brand starts and formulary tier placement ahead of pharmacy price.
Name the node where arithmetic is exact but attribution is not: the risk-adjustment denominator. Dividing by the population mean prospective_risk_score is exact arithmetic, but the score is built from coded diagnoses, so more complete coding raises the denominator and lowers risk-adjusted PMPM with no change in care delivered. Carry raw PMPM, the mean risk score trend and a coding-intensity measure such as diagnoses per member-year beside it, permanently.

Worked solution (45 min)

Write the base identity and the three-term change decomposition on paper first, and confirm the terms sum to the total change on a two-category toy example before touching real data.
Specify the preparation rules that precede any aggregation: dedup medical_claim_line to the surviving version per claim_id using claim_version and frequency_code, drop both rows of every reversed pharmacy pair, and build member-months from coverage spans with partial months counted fractionally.
Draw the medical branch and the pharmacy branch to leaf level, writing the identity beneath each split rather than only the metric name.
Write the generic-dispensing split as the exact price identity and confirm it reproduces the observed allowed per 30-day script.
Attach one leading indicator and one named owning function to each leaf.
Write the caveat on the risk-adjustment node, including the two companion series that must be published with it.

Expected result: A four-level tree rooted at risk-adjusted allowed PMPM, splitting by service category into utilization per member-month and allowed per unit, with an inpatient sub-split on DRG case mix, an ambulatory sub-split on place_of_service_code, and a pharmacy branch on 30-day equivalents whose price term is split exactly by generic dispensing rate; each leaf carries a named owner and a leading indicator; and the risk-adjustment denominator is flagged as exact arithmetic with unreliable attribution, published alongside raw PMPM, the mean risk score trend and a coding-intensity series.

Verify it yourself

Reconcile the leaves back to the root: the summed category contributions must equal total allowed PMPM to within rounding, and the three change terms must sum to the observed period change.
Recompute total allowed with and without version dedup. Any gap is the double-counted adjusted claims the tree would otherwise attribute to real utilization.
Verify the pharmacy price identity numerically: generic dispensing rate times generic price plus its complement times brand price should reproduce the observed blended allowed per 30-day script.
Hold care constant and simulate a 5 percent rise in mean prospective_risk_score. Risk-adjusted PMPM should fall by roughly the same proportion, which is the attribution problem made visible.

Follow-up

Utilization is flat and allowed per unit is flat, yet PMPM rose 4 percent. Name every place the increase can be hiding.
Your member-month denominator came from coverage spans that overlap across products for some members. What does that do to PMPM, and how do you decide whether to deduplicate?
Which leaf would you refuse to put on a team's scorecard, and why?

First-pass denial rate doubled in one received week

Medium · pipeline failure · segment cuts · denominator definition

First-pass denial rate jumped from 6.1 to 11.8 percent in a single received_date week and stayed at the new level. Nothing in the adjudication rules changed that week. You have medical_claim_line: claim_id, claim_version, frequency_code, claim_status, denial_reason_code, member_id, received_date, adjudicated_at, billing_provider_npi. The metric counts lines whose first adjudication returns 'denied', over lines received in the window at claim_version = 1. Find the cause, say whether denial behaviour actually changed, and deliver the corrected series plus the query change that fixes it.

Approach

Cut the excess by denial_reason_code before anything else. A behaviour change spreads across codes; an upstream failure concentrates in one. If a single eligibility-related code carries nearly all of the 5.7 point excess, the claims are fine and the eligibility data was not.
Join the denied lines to member_enrollment and compare the span's created_at to the claim's received_date. Lines denied for members whose coverage span loaded after the claim arrived are a load-ordering failure, not a coding failure.
Check resolution downstream: take those claim_ids and look for a later version with frequency_code 7 that adjudicates to 'paid'. A high resubmit-and-pay share confirms the denial was transient and recoverable.
Verify the denominator is not the mover. Count lines at claim_version = 1 by received_date and check the weekday mix, because submission batching makes received_date strongly day-of-week patterned and a short or holiday week shifts the denominator without changing behaviour.
Confirm runout: the metric should be held until 60 days of adjudication have passed, so check that the comparison weeks are equally mature rather than comparing a settled week to a fresh one.
Restate the series with the affected reason code broken out as its own line, and change the query to carry denial_reason_code into the reported breakdown rather than only the headline rate.

Follow-up

The affected lines all get paid on resubmission. Is the first-pass denial rate then wrong, or right and uninteresting?
How would you monitor for this class of failure without waiting for someone to notice a chart?
What changes if the excess is spread evenly across denial_reason_code instead of concentrated?

Suggested practice checklist

For a candidate whose interviews will centre on A/B testing, metric movement and causal claims. Design comes before arithmetic, arithmetic before analysis, and the week ends by rehearsing the readout rather than the derivation.

Day 1 — Design one test end to end on paper

Take a single feature change and write the full design: randomization unit, the exact point of exposure, the primary metric with its grain, guardrails, allocation, planned duration, and the decision rule committed before any data exists.
Write why the randomization unit must sit at or above the level where treatment can spill over, and give one case where user-level randomization is still contaminated (shared accounts or devices, or two participants in the same marketplace).
State in advance what you will do if the primary metric is flat while a secondary metric is significant.

Deliverable: A one-page test design with a decision rule written before launch.

Day 2 — Power arithmetic until it is automatic

Compute required sample size per arm for a binary metric with the normal approximation, n is approximately 2 times (z for alpha/2 plus z for power) squared times p(1 minus p) divided by delta squared, for baselines of 2, 10 and 40 percent at a 5 percent relative lift, and note that for a fixed relative lift the requirement falls as the baseline rises because delta grows proportionally with p.
Redo the calculation for a continuous metric using variance in place of p(1 minus p), and show why a heavy-tailed quantity such as revenue per user needs either far more traffic or a capped version with a stated cap.
Convert one of the results into weeks given a weekly eligible traffic figure, then list the two honest ways to shorten it (accept a larger detectable effect, or reduce variance) and write why quietly lowering the power target is a decision to miss more real wins, not a speedup.

Deliverable: A small script or sheet that maps baseline, minimum detectable effect, alpha and power to sample size and weeks, cross-checked against a published calculator.

Day 3 — Variance and the unit-of-analysis problem

Take a ratio metric whose denominator is not the randomization unit (clicks per session, randomized by user) and compute the standard error twice, once naively at session level and once by the delta method or a user-level bootstrap, then record how much the naive version understates it.
Implement CUPED on simulated data: choose a pre-period covariate X measured before assignment, estimate theta as Cov(Y, X) divided by Var(X), and analyse Y minus theta times (X minus its mean) in place of Y. Confirm the variance of the adjusted outcome equals the raw variance multiplied by one minus the squared correlation between Y and X, so a correlation of 0.45 removes about 20 percent of the variance and not 80.
Now run that simulation a few hundred times and confirm the adjusted effect estimate is unbiased for the same effect rather than numerically identical to the raw one. Within any single run the two differ, sometimes by a large fraction of the true effect, because the two arms' pre-period covariate means never coincide exactly in a finite sample; they agree in expectation, which is the property that matters and the one to state out loud.

Deliverable: A notebook showing the adjusted estimator with a measurably smaller variance than the raw one, plus a repeated-simulation table showing the two estimators agreeing on average while differing run by run.

Day 4 — Validity threats you can actually test for

Run a sample ratio mismatch check as a chi-square goodness-of-fit test against the intended allocation, and write the three causes you would chase first (assignment logged before exposure, an arm-specific redirect or load failure, bot filtering applied asymmetrically).
Simulate peeking: generate A/A data, test daily at alpha 0.05 across 14 looks, record the inflated false positive rate, then apply an alpha-spending boundary or commit to a fixed horizon and confirm the rate returns to nominal.
Write how you would separate a novelty effect from a durable lift using the treatment effect plotted against days since first exposure, and what shape would change your recommendation.

Deliverable: One table showing the peeking false positive rate before and after correction, plus a written SRM triage list.

Day 5 — When randomization is not available

Write the identifying assumption for difference-in-differences (parallel trends in the absence of treatment), then plot pre-period trends for two candidate control groups and justify rejecting one of them.
Design a switchback test for a change where user-level randomization would leak across participants, choosing a time-block length against the carryover you expect and saying how you would detect carryover in the data.
List what an interrupted time series or a synthetic control buys you and the one thing neither can rule out: an unobserved shock that coincides with the launch.

Deliverable: A one-page memo recommending a single quasi-experimental design and naming its weakest assumption explicitly.

Day 6 — The readout query

Write the assignment-to-exposure join that returns exactly one row per unit per experiment, and handle units appearing in both arms by excluding and counting them rather than silently keeping one.
Compute the per-arm metric, its variance and the relative lift with a confidence interval in SQL, then reproduce the identical numbers in a notebook as a cross-check.
Add a segment breakdown and write the sentence that keeps it from being p-hacking: segments declared in advance, everything else reported as exploratory and corrected for multiplicity.

Deliverable: A single query that outputs the full readout table, matched to a notebook recomputation.

Day 7 — Present it to someone who will not read the appendix

Give a 10-minute readout of a real or simulated experiment in the order decision, number, uncertainty, caveat.
Have your listener ask "can we ship it" in the case where the primary is flat and a guardrail moved, and answer with a recommendation rather than a request for more data.
Rewrite your opening line so the recommendation lands before any methodology.

Deliverable: A one-page readout whose first line is the recommendation.

Ambiguous requests are the test

Most of the questions in this section reduce to one thing: can you be handed a vague request and come back with something useful? Prepare an example where the ask was underspecified, you chose an interpretation, and you said out loud which interpretation you chose. Describing how you narrowed the question matters more than the technique you eventually used.

Why do you want to work as a Data Scientist specifically in the pharmaceutical industry and at GSK?

Medium · behavioural and stakeholder questions

Approach

Name the disagreement or constraint, and how you resolved it with evidence.
Pick a story where you drove the decision, not one where you observed it.
Quantify the outcome, including what you would not claim credit for.

Follow-up

What did you decide not to do, and why?
How did you know the outcome was caused by your change?

Tell me about a time you developed a technical project and encountered a major data quality issue. How did you identify and resolve it?

Medium · behavioural and stakeholder questions

Approach

Quantify the outcome, including what you would not claim credit for.
Close with what you would do differently, concretely.
Name the disagreement or constraint, and how you resolved it with evidence.

Follow-up

How did you know the outcome was caused by your change?
What did you decide not to do, and why?

Make an honest case for your own analytical impact

Hard · impact attribution · counterfactual · self-assessment

You are asked to account for a year of your work. Constraint: you may only use measures that existed before you knew the outcome, and you must include one project that moved nothing. The portfolio contains a measure specification that changed how a contract settles, a predictive model that shipped, a self-serve dashboard, and an evaluation that killed a proposed programme. Deliverable: how you attribute value to each, which of them you claim credit for and which you do not, and the one number you would accept being judged on next year.

Approach

The probe is whether you can reason counterfactually about your own work, which is the same discipline this domain demands of you for programmes, applied where it is uncomfortable.
State the counterfactual for each item: what would have happened without you. A measure specification that changed how a contract settles has a dollar counterfactual you can compute. A dashboard's counterfactual is analyst hours, which is real and small, and attaching a revenue number to it is where these accounts usually turn into fiction.
Treat the killed programme as your strongest item and say why. Avoided spend is real value, it is attributable because the decision turned on your evaluation, and it is the one item nobody else will count on your behalf.
Report the shipped model the way you would report anyone's model: calibration in the subgroups that matter, adoption by the people it was built for, and whether any decision changed. A model that is live but ignored has zero impact, and saying so is the credibility test in this answer.
Name the failed project with its mechanism and its cost, then say what earlier signal you would watch to kill it sooner. Time to kill is a legitimate impact measure.
Choose next year's number so that it can go down. A metric that only ratchets upward is not a commitment.

Follow-up

Two other people also claim the contract measure. How do you split it?
Your model is live and used, but the decisions it informs would have been the same without it. Is that impact?
What is the strongest argument that your year had less impact than you just described?