Preguntas de entrevista investigadas, descripción del proceso y señales de dificultad en Affirm, compiladas por el equipo de investigación de Primly.
6 experienciasDificultad 3.5/5FinTech / BNPL
Senior Data Scientist (Risk Analytics)
virtual
· Dificultad 4/5
Candidates report a recruiter screen covering domain alignment for consumer lending, experimentation and modeling background, and collaboration style across product, engineering, and risk stakeholders. The first technical round is often a video interview with a data scientist or analyst that includes SQL and analytics case questions tied to underwriting performance, funnel drop-off, loss rates, and cohort behavior. Many interview loops include a modeling or decision-science interview that asks candidates to reason about credit risk, bias and fairness considerations, model monitoring, and trade-offs between approval rate, losses, and customer experience. A take-home is less consistently mentioned than live casework, but when used it tends to be a compact analysis or a metrics and experiment design exercise with an emphasis on communicating assumptions and limitations. Final rounds typically add cross-functional interviews with risk product partners or engineering to test how candidates translate analysis into policy or model changes and how they operate in a regulated, audit-aware environment; overall timelines are commonly described as 3 to 6 weeks.
Given a dataset of loans with outcomes, what features would candidates create to predict default risk at checkout, and how would they avoid target leakage?
If approval rates increased after a policy change but losses also rose, how would candidates diagnose whether the change caused the loss increase versus seasonality or merchant mix shift?
Design an experiment to test a new underwriting threshold or model, including guardrails that protect consumers and the business, and metrics candidates would monitor week one versus month one.
Describe a time candidates disagreed with a product or business partner about a metric or decision. How did they resolve it and what did they ship as a result?
How would candidates evaluate and mitigate unfair impact in a credit model, and what would they communicate to stakeholders about trade-offs and compliance constraints?
Customer Support Specialist
virtual
· Dificultad 2/5
Candidates report starting with a recruiter phone screen focused on availability, shift expectations, remote or site requirements, and baseline customer service experience in regulated environments. The next step is typically a hiring manager video interview that tests de-escalation, written and verbal communication, and comfort handling identity, payments, disputes, and installment plan questions specific to BNPL. Many processes include a work sample, such as drafting customer email responses, documenting a case in a ticketing-style format, or role-playing a live chat scenario where policy and empathy both matter. A final round often adds a second operations leader or quality assurance partner to probe consistency, policy adherence, and how candidates handle ambiguous edge cases, including fraud or sensitive financial hardship. End-to-end timing is commonly described as about 2 to 4 weeks depending on scheduling and background check requirements.
A customer says their Affirm payment is overdue because they never received the merchandise. How would candidates investigate and respond while following policy and maintaining empathy?
What steps would candidates take to verify identity before discussing account details, and what would they do if the customer cannot pass verification?
Describe a time candidates had to calm an angry customer while still enforcing a rule they did not like. What language did they use and what was the outcome?
During a live chat, a customer claims an installment loan was opened without their authorization. How would candidates triage for potential fraud and escalate appropriately?
Affirm emphasizes transparency and responsible lending. How would candidates explain a denial or a limit decision in a way that is clear and non-defensive?
Credit Risk Manager
virtual
· Dificultad 4/5
Recruiter screen, technical phone (stats + credit fundamentals), virtual onsite with case study on a credit-portfolio question, behavioral, and final with director. Credit Risk at Affirm manages billions in loan exposure.
Walk me through analyzing a credit-portfolio performance.
Tell me about an underwriting model improvement.
Describe partnering with product on a credit-policy change.
How do you balance loss rates with origination volume?