Interview guides · Amex

Amex Data Scientist Interview Questions & Guide 2026

A two-week prep plan for Amex data scientist candidates: write your answer before reading the solution, revisit misses three days later.

The 60-second brief

As a Data Scientist at Amex, you play a pivotal role in driving data-driven decisions that shape global payment products, credit risk models, marketing strategies, and fraud detection systems. You operate at the intersection of complex financial data, advanced machine learning, and core business strategy. Your daily work directly impacts millions of cardmembers worldwide by securing transactions, optimizing credit limits, and personalizing financial services.

How much statistics you need depends on the flavour of the seat. Experiment-facing work wants you deep enough to notice that repeated looks at accumulating data inflate the false positive rate of a fixed-sample test; modelling-facing work wants estimation and honest uncertainty intervals.

Focus areas

Read vintage curves, not blended portfolio averages
Separate authorization, settlement and dispute outcomes cleanly
Decompose expected loss into PD, LGD, EAD

Overview

The scope of this role at Amex is uniquely challenging due to the massive scale and velocity of transaction data. You will collaborate closely with product managers, data engineers, and business leaders to build predictive models, design experimentation frameworks, and translate complex algorithms into tangible business value. Whether you are developing gradient-boosted fraud classifiers, evaluating credit risk portfolios, or designing product metrics, your insights directly influence the bottom line of a global financial leader.

Expect a fast-paced and intellectually rigorous environment where theoretical knowledge must seamlessly translate into real-world applications. Interviewers will look for your ability to balance technical sophistication with practical business acumen. If you thrive on solving high-impact problems in a data-obsessed culture, this role offers an exceptional platform to scale your career.

The Amex interview process

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

Recruiter Screen

A screening call is a matching exercise run by someone who will not evaluate your statistics. They are checking that the work described on your resume is work you personally did, and that its scope matches the level the role is written for. Logistics get settled in the same half hour so nobody spends an interviewer's afternoon on a mismatch. The answer that fails is the one narrated in the plural. If every sentence is 'we built' and 'the team decided', there is nothing specific to write down about you. Name the piece that was yours, the decision you made inside it, and what changed after.

What to demonstrate

Whether the ownership implied by your resume survives one round of follow-up about who actually did which part
Whether your described scope (data size, stakeholders, what shipped) matches the seniority the role is written at
Whether timeline, location and compensation expectations make the rest of the loop worth scheduling

How to prepare

Rewrite your top three resume bullets in the first person singular, each with the decision you made and what moved afterwards, then say them out loud once so the 'we' does not return under pressure
Attach one number to each project: the baseline, the change, and the window it was measured over. Where impact was never measured, say that plainly rather than inventing a figure
Settle your compensation range before the call and give it as a range with a reason behind it, such as current total comp or a competing timeline, instead of deflecting the question twice

Technical Assessment

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

Interview Panels

Where a loop includes a partner from outside the data team, that conversation usually carries the same weight as the technical ones and gets the least preparation. The person opposite you will not follow a derivation and does not need to. They are working out whether having you involved would make their decisions better or slower. The failure mode is not being too technical. It is answering a question about a decision with a description of your method, leaving the translation to them. What they carry into the debrief is the sentence you handed them, not the analysis underneath it.

What to demonstrate

Whether a statistical result arrives as something the partner could act on, with the one caveat that would change their decision kept and the rest left out
Whether you can state what you need from their side, in their terms: instrumentation that does not exist yet, a definition they own, or a holdout they have to agree to
Whether uncertainty is given as a range someone can plan against, rather than as hedging that invites them to ignore the result
Whether you ask what decision is actually on the table before explaining anything

How to prepare

Take a result you know well and write the version for someone who stops reading after one sentence, then the three-minute version, and check the short one is not the long one with the qualifications stripped out
For a past project, list everything you asked a non-technical partner for and how you phrased it, then rewrite each ask so it names what goes unmeasured without it
Practise saying where a result does not apply, out loud, in one sentence that a partner could repeat accurately to someone else

Deep Dive Interviews

An added round often puts you in front of someone outside the core hiring team: a partner engineer, a product owner, a domain expert, sometimes a more senior manager. The question they are really asking is not whether you can do the work but whether they would trust a number that came from you. That changes what a good answer looks like. Lead with what the decision cost and what it changed, keep the method available but not central, and be plain about the limits of your evidence. Overstating a result is the fastest way to lose this round.

What to demonstrate

Whether you can explain a technical choice to someone who will never read your code, without either flattening it into nothing or hiding inside jargon
Honesty about evidence strength: what the analysis establishes, what it only suggests, and what it cannot say at all
How you take disagreement, specifically whether you update on a good objection, hold your position with reasons, or fold on contact

How to prepare

Write the two-sentence version of your most technical project for a non-specialist, then check that neither sentence needs a method name to make sense.
For one result you are proud of, write the strongest objection someone could raise and a response that concedes the part of it that is correct.
Prepare one decision that turned out to be wrong: how you found out, what it cost, and what you changed afterwards. A senior cross-functional interviewer asks for this more often than a technical one does.

Where candidates lose points

Assuming a model is fair because protected attributes are not among its inputs

Postcode, device, tenure, income proxies and even transaction patterns correlate with protected characteristics, so a model can produce a disparate outcome without ever reading the attribute. Credit decisions additionally carry an explainability obligation in many jurisdictions, since a denial has to be accompanied by its principal reasons, which constrains model form and feature engineering rather than being a reporting afterthought. Treating fairness testing and reason-code generation as design constraints from the first model version is far cheaper than retrofitting them to a deployed one.

Averaging delinquency across a book that is growing

A loan three months old cannot be 90 days past due, so a portfolio with many recent originations reports a low blended 90+ rate purely from age mix. The blended rate falls fastest exactly when originations grow fastest, which is precisely when credit quality most needs watching, so the metric moves in the reassuring direction during the riskiest period. Only comparisons at equal months on book are valid, which is what a vintage or roll-rate view enforces.

Never asking what decision the analysis will inform

Open with who makes the decision, what the options are, and by when. The answer determines the precision you need, the segments worth cutting, and whether an observational read suffices or an experiment is required.

Building features from data that postdates the prediction time

Check every feature against the timestamp at which the model would actually score, and drop anything computed from a window that includes or follows the label event. For a forecasting use case, split train and test by time rather than at random, and split by entity when the same entity recurs.

Questions & practice

1–2 sentences introducing the category and what it tests.

Medium · statistics and probability

Approach

Say what the estimate is of, and over what population it generalises.
Write down the assumption the method needs before you use the method.
Quantify uncertainty explicitly rather than reporting a point estimate alone.

Follow-up

How would you explain this result to someone who does not know statistics?
What sample size would you need to detect an effect half this size?

1–2 sentences introducing the category and what it tests.

Medium · machine learning and modelling

Approach

Set a baseline first, so any model has something honest to beat.
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

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?

Bullet list of realistic information: What is the difference between R…

Medium · machine learning and modelling

Bullet list of realistic information: What is the difference between Random Forest and Gradient-Boosted Decision Trees, and when would you use one over the other? How do you handle extreme class imbalance in fraud detection models? Explain how model interpretability techniques like SHAP or LIME work in production environments. What metrics would you use to evaluate a credit risk scoring model beyond accuracy?

Approach

Set a baseline first, so any model has something honest to beat.
Check what information would not exist at prediction time, and exclude it.
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?
Where could label leakage enter this setup?

Estimate a delinquency roll-rate matrix and project twelve months

Hard · roll rates · markov chain · survivorship

fct_loan_performance_monthly gives loan_id, as_of_month_end, months_on_book, delinquency_bucket, charge_off_flag, prepaid_in_full_flag and restructured_flag. Build a month-to-month transition matrix over the five delinquency buckets plus absorbing charged_off and prepaid states. Loans that stop appearing must be routed to an absorbing state rather than dropped. Project the current book forward 12 months by repeated matrix multiplication and report the projected share reaching charge-off. Handle restructured_flag explicitly, and name one place the Markov assumption fails on this data.

Approach

Build consecutive month pairs per loan by shifting as_of_month_end within loan_id, then verify the shifted value is exactly one month later. A gap is not a transition, it is an exit you have not resolved yet.
Resolve exits before counting anything. A loan whose last row carries charge_off_flag moves to charged_off, one carrying prepaid_in_full_flag moves to prepaid, and one that disappears with neither is a data question to raise rather than silently discard, because discarding it is survivorship that inflates every cure rate.
Count pairs into a 7 by 7 matrix and row-normalise. Assert every row sums to one and the two absorbing rows are the identity; a row that does not sum to one means exits were dropped.
Decide and state the restructure rule. Restructuring resets days_past_due, so a dpd_60_89 to current move on a restructured loan is not a cure. Either give restructured loans their own state or carry the pre-restructure bucket, but do not let that move land in the cure cell.
Project by taking the current bucket distribution as a row vector and multiplying by the matrix twelve times. Report the charged_off entry, and report it again from an all-current starting vector so the reader can see how much of the projection comes from loans that are already delinquent today.
State the homogeneity failure plainly: transition rates depend strongly on months_on_book, so one pooled matrix applied to a book with a young mix understates early-life delinquency. If the mix is moving, estimate separate matrices by seasoning band.

Worked solution (45 min)

Sort by loan_id and as_of_month_end, shift to form (from_state, to_state) pairs, and flag pairs whose month gap is not exactly one.
For each loan's final row, assign the absorbing destination from charge_off_flag or prepaid_in_full_flag, and list loans that vanish with neither as an exception count to report.
Apply the restructure rule, then build the 7 by 7 count matrix with a cross-tabulation over ordered state categories and row-normalise it.
Assert row sums equal one and absorbing rows are the identity, then take the current month's bucket distribution as a row vector.
Multiply twelve times, report the charged_off component, and repeat from an all-current vector for comparison.

Expected result: A 7 by 7 row-stochastic matrix with identity rows for charged_off and prepaid, a current-to-current diagonal typically above 0.95, roll rates rising with bucket depth, and a 12-month projected charge-off share of a few percentage points from an all-current start and materially higher from the actual book.

Verify it yourself

Every row sums to one within floating-point tolerance, and both absorbing rows are exactly the identity.
No mass sits in cells that skip a bucket upward, such as dpd_1_29 to dpd_60_89 in one month; if there is, the pairs are not actually consecutive months.
The projection from an all-current start is strictly below the projection from the real book distribution.
Backtest by estimating on months up to t minus 12 and comparing the projected distribution with the realised one at t; a systematic optimistic gap points at dropped exits or at seasoning, not at noise.

Follow-up

How would you validate the projection against what actually happened, and over what window?
The cure rate out of dpd_30_59 rose five points last quarter. What are the candidate explanations and how would you separate them?
When would you prefer a vintage curve to a roll-rate projection, and why?

Bullet list of realistic example questions: How would you figure out w…

Medium · sql and data manipulation

Bullet list of realistic example questions: How would you figure out whether a transaction is fraudulent? How to predict whether a customer will call us in the near future for any query? How does the credit market work and what are our core revenue streams? How would you evaluate the quality of a recommender system using online versus offline metrics? Case study on marketing data science and campaign attribution.

Approach

Compute rates by summing numerator and denominator separately, never by averaging rates.
Say which table is the grain you start from, and join outward from it.
State the window function and its partition and ordering out loud before writing it.

Follow-up

What breaks if events arrive late or out of order?
How does the query change if the join becomes one-to-many?

1–2 sentences introducing the category and what it tests.

Medium · sql and data manipulation

Approach

Say which table is the grain you start from, and join outward from it.
Handle the rows that do not match: a LEFT JOIN with a NULL check is usually the question.
Compute rates by summing numerator and denominator separately, never by averaging rates.

Follow-up

What breaks if events arrive late or out of order?
How would you verify this result without re-running the same query?

Customers with no credit application, avoiding the NOT IN trap

Medium · null semantics · anti-join · scd type 2

Count current customers who have never submitted a credit application, broken out by segment. dim_customer is a slowly changing dimension type 2, so restrict to is_current = true, kyc_status = 'verified' and closed_at null. In fct_loan_application, customer_id is null for applicants who were not customers when they applied. Write the anti-join, return segment and customer_count, and state in one line what NOT IN (SELECT customer_id FROM fct_loan_application) returns against this table and why.

Approach

Pin the dimension to one row per customer first: is_current = true already guarantees that, but say so out loud, because forgetting it multiplies every count by the number of attribute versions a customer has accumulated.
Write the anti-join as NOT EXISTS with a correlated predicate on customer_id, which evaluates per row and is unaffected by nulls anywhere in the applications table.
Name the failure explicitly: NOT IN against a nullable column compares each candidate to a set containing NULL, the comparison yields UNKNOWN rather than TRUE, and the whole predicate is therefore never satisfied, so the query returns zero rows.
If NOT IN is required for some reason, add WHERE customer_id IS NOT NULL inside the subquery, which restores the intended semantics, and note that a LEFT JOIN with an IS NULL filter is equally safe.
Group by segment and sanity-check the total against the unfiltered current-customer count minus the count of distinct applying customers.

Worked solution (20 min)

SELECT segment, COUNT(*) FROM dim_customer c WHERE c.is_current AND c.kyc_status = 'verified' AND c.closed_at IS NULL.
Add AND NOT EXISTS (SELECT 1 FROM fct_loan_application a WHERE a.customer_id = c.customer_id).
Group by segment and order by the count descending.
Run the NOT IN variant alongside it and record that it returns zero rows, then run it again with IS NOT NULL added to the subquery and confirm the counts match the NOT EXISTS version.

Expected result: One row per segment present in the current dimension (retail, sme, corporate), each with a non-zero customer_count. The NOT EXISTS and the null-guarded NOT IN forms agree exactly; the unguarded NOT IN form returns no rows at all.

Verify it yourself

COUNT(DISTINCT customer_id) equals COUNT(*) on the filtered dimension, which confirms is_current really collapsed the type 2 versions.
The total across segments plus COUNT(DISTINCT customer_id) from fct_loan_application restricted to current verified open customers equals the full current verified open customer count.
Swapping NOT EXISTS for the unguarded NOT IN changes the result from a populated set to an empty one, which is the whole point of the exercise.

Follow-up

Rewrite it as a LEFT JOIN with IS NULL and say when you would prefer that form to NOT EXISTS.
How does the answer change if you want customers who never applied as of a historical date rather than today?
The applications table has 40,000 rows with a null customer_id. What are those rows, and are they a data quality problem or a product fact?

1–2 sentences introducing the category and what it tests.

Medium · product sense and metrics

Approach

Restate the decision this analysis has to support, and who acts on the answer.
Decompose the metric into the rates that drive it, and say which one you would check first.
State what result would change your recommendation, so the answer is falsifiable.

Follow-up

Which segment would you cut first, and what would that rule out?
How would you detect that the metric is being gamed rather than genuinely improving?

1–2 sentences introducing the category and what it tests.

Medium · a/b testing and experimentation

Approach

Say whether units interfere with each other, and switch design if they do.
Name the randomisation unit first; it decides the variance and what the test can detect.
Name the guardrails that would stop a launch even on a positive primary result.

Follow-up

How would you handle interference between treated and control units?
What would you conclude if the result is positive but the test is underpowered?

Measure an underwriting automation programme without hiding cost

Medium · metric gaming · swap set · operations metrics

An operations programme aims to cut manual underwriting. The proposed metric is manual review rate: decided_by = 'manual' over applications with decision in ('approve','decline') in fct_loan_application. The team controls the referral rules recorded in policy_rule_hits. Explain how they can halve that metric in one release while making the book worse, then design the readout that catches it: primary metric, the cuts it must always be reported in, and the guardrails. Say what evidence you would require before declaring a batch of referrals safe to auto-decide.

Approach

Name the exploit precisely. Referrals can be converted to automatic declines by a rule change, which removes them from the numerator at no operational cost while they stay in the denominator, so the rate falls and applicants a human would have approved are rejected instead. Auto-approving the same population moves the metric identically with the mirror-image failure mode.
Refuse the bare rate as a primary. Make it conditional on outcomes: manual review rate measured at a held-constant approval rate within each model_pd_12m band, so a reduction is creditable only when the decisions themselves did not move.
Require fixed cuts on every report: by model_pd_12m band and by the rule identifier inside policy_rule_hits. A single blended operations number cannot be decomposed after the fact, and the decomposition is the entire diagnostic.
Set guardrails on the things the exploit disturbs: approval rate within each score band, the override rate where manual decisions disagree with the automated recommendation, and the vintage 90+ rate at months_on_book = 12 for the newly auto-decided population against the referrals of the prior regime.
Fix the evidence bar for retiring a referral rule: a shadow period where the rule fires and the case is still worked manually, recording what the automation would have decided, what the human decided, and, for the disagreement group alone, the realised performance. Agreement on the easy majority establishes nothing about the disputed minority.
State the conflict honestly rather than designing around it. The referral population exists because the model is least certain there, so automating it trades loss for cost by construction, and the readout's job is to price that trade rather than to make it disappear.

Worked solution (30 min)

Compute manual review rate for a recent month overall, then by model_pd_12m band, then by rule identifier extracted from policy_rule_hits.
For each rule, compute the approve share and decline share of the applications it touches, so the rules whose removal would actually change decisions are identified before anyone proposes removing them.
Simulate the exploit: reclassify one high-volume referral rule's applications as automatic declines, then recompute the blended rate and the band-level approval rates side by side.
Write the readout spec: the primary with its conditioning, the two mandatory cuts, the three guardrails, and the shadow-period evidence bar.

Expected result: A table in which the simulated rule change lowers the blended manual review rate materially while approval rate in the affected model_pd_12m bands falls at the same time, together with a readout spec that makes it impossible to publish the first effect without the second.

Verify it yourself

The simulated change lowers the manual review rate and the band-level approval rate together. If band approval rate is unchanged, that rule was not binding and its removal is genuinely free.
Rule-level counts reconcile to the blended rate once applications hitting several rules are counted once. A jsonb array of rule identifiers is the usual place this double-counts.
The newly auto-decided population and the prior manual referrals are compared at equal months on book, not at equal calendar dates.

Follow-up

The override rate falls to near zero after the change. Is that good?
How long must the shadow period run, given the outcome matures at twelve months on book?
The rule fires on 200 applications a month. What can and cannot be established at that volume?

Approval rate fell but approved value did not

Easy · approval rate · retry chains · denominators

Over ten days the count-weighted 7-day approval rate on fct_payment_authorization fell from 91 to 86 percent, while captured value in the reporting currency is flat. Available columns: auth_id, card_token_id, merchant_id, mcc, channel, requested_at, amount_minor, transaction_currency, auth_result, decline_reason_code, risk_score, is_reversal, parent_auth_id, captured_at, captured_amount_minor, issuer_country. In twenty minutes, decide whether approved value is actually at risk, and hand back a corrected rate together with its denominator and every exclusion written down.

Approach

Restate the denominator before querying anything. The current one counts every row, so reversals (is_reversal = true), incremental authorizations (parent_auth_id not null) and zero-amount verification attempts (amount_minor = 0) are all sitting in it.
Chart numerator and denominator separately by day. If approved counts are flat and total attempts rose, the rate moved because the denominator grew, which is a different investigation from a rule change and points at a different owner.
Group declines by decline_reason_code and merchant_id. Retry-driven inflation concentrates in a few soft decline codes at a few merchants; a genuine policy change spreads across merchants within one code family.
Collapse retry chains: partition by (card_token_id, merchant_id, amount_minor), keep one attempt per 15-minute window taking the best outcome, and recompute both count-weighted and dollar-weighted rates on the collapsed set.
Convert amount_minor to one reporting currency using each currency's ISO 4217 exponent before any dollar weighting, then cut by channel and issuer_country to confirm nothing is hiding underneath a flat total.
Report both rates side by side with the exclusion list attached, and state which definition the alert should have been built on.

Follow-up

How would you pick the retry-collapsing window when merchants retry on different schedules?
Flat captured value could itself be hiding a mix shift. How do you rule that out?
What monitor would have caught denominator inflation on the day it started, rather than ten days later?

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.

Getting an analysis acted on

Work that nobody used is a common and unflattering pattern in data careers, and interviewers probe for it. Have a story about an analysis that changed a decision, and be specific about how you got it in front of the person who could act. Also have one about work that went nowhere, with your reading of why.

1–2 sentences introducing the category and what it tests.

Medium · behavioural and stakeholder questions

Approach

Quantify the outcome, including what you would not claim credit for.
Name the disagreement or constraint, and how you resolved it with evidence.
State the situation in two sentences and spend the rest on your reasoning.

Follow-up

What would you do differently if you ran that project again?
How did you know the outcome was caused by your change?

Defend a vintage finding that contradicts the portfolio dashboard

Medium · vintage analysis · mix shift · stakeholder pushback

The lending dashboard shows blended 90-plus days-past-due falling for four consecutive quarters while originations grew 60 percent. Using fct_loan_performance_monthly, you build a vintage view keyed on origination_month by months_on_book and find the three most recent vintages are worse than their predecessors at the same age. The business lead presents that dashboard weekly and pushes back hard, suggesting you picked favourable cohorts. You get one meeting and the vintage table. Present the finding so it survives the cherry-picking objection and ends in a decision.

Approach

Reconcile before you contradict: show that aggregating your vintage table along the calendar diagonal reproduces the published blended series, so the disagreement is about age mix rather than about data quality.
Make the mechanism arithmetic rather than rhetorical: a loan cannot reach 90 days past due before it is 90 days old, so rapid origination growth shifts weight onto young months-on-book where the rate is structurally near zero.
Show every vintage rather than a selected pair, all indexed at months_on_book equal to 12, with cohort sizes printed beside each curve so nobody can claim the divergence rests on a thin cohort.
Handle restructuring explicitly, because restructured_flag resets days_past_due: count each loan on its worst pre-restructure state, or recent vintages will look better than they are.
Close on the decision rather than the chart: state what the divergence implies for the cutoff or the channel mix, and state in advance what evidence would make you withdraw the claim.

Follow-up

Two cohorts differ at month 12. How do you separate a seasoning effect from a genuine credit-quality effect?
Someone argues the recent vintages are simply a broker-channel mix shift. How do you test that, and what would confirm it?

Turn a one-line fraud-number request into a scoped brief

Easy · scoping · metric definition · denominators

A stakeholder messages: what is our fraud rate, and is it going up? You have fct_payment_authorization, fct_card_dispute and dim_customer. At least four defensible answers exist: count-weighted or value-weighted, attributed to the transaction month or to the dispute filing month, and gross or net of recoveries and successful representments. You get one reply before someone else produces an uncaveated number. Write that reply: the clarifying questions you ask, the single default you will produce if nobody answers, and what the default excludes.

Approach

Establish the decision behind the question first, because a risk-rule change, a board number and a merchant contract negotiation need different denominators, and asking which one is not stalling.
Offer a short menu rather than an open question: a stakeholder can choose between two named options but cannot specify a denominator from scratch.
Commit to a default so the reply is useful even if nobody answers, for example net fraud loss in basis points of settled volume, attributed to the requested_at month, matured months only.
State the exclusions in the same breath as the default: non-fraud dispute categories, transaction months with less than 120 days of maturity, and first-party abuse that arrives coded as consumer_dispute.
Give a delivery time for the default and a longer one for the fuller cut, so the choice between them carries a visible cost.

Follow-up

They come back wanting it by merchant for a contract negotiation. What changes in the definition and in the maturity rule?
How would you separate first-party abuse from third-party fraud in this data, and what would you refuse to conclude from the split?