Build a Personalized Ranking System for Financial Products (App Home Page)

Context

You are the first Data Scientist partnering with a PM to build an end-to-end personalized ranking of financial products (e.g., loans, credit cards, deposits, investments) on the home page for a SoFi-like fintech. You must balance short-term engagement with long-term, risk-adjusted business value while meeting strict eligibility and compliance constraints.

Tasks

1. Objective(s) and Objective Function
   • Define the primary objective (e.g., risk-adjusted LTV over 30 days) and secondary objectives (CTR, application starts, approvals).
   • Write an explicit objective function that trades off short-term clicks vs. long-term funded accounts, subject to eligibility/compliance constraints.
2. Learning-to-Rank Choice
   • Choose pointwise, pairwise, or listwise learning-to-rank and justify.
   • Specify the loss function and any calibration you’ll apply.
3. Features and Cold Start
   • List user, product, and context features.
   • Explain how you’ll handle cold start for new users and new products.
4. Bias and Feedback Loops
   • Address position bias and selection bias.
   • Describe counterfactual logging and IPS or doubly robust estimators you’ll use.
5. Personalization + Business Rules
   • Blend personalization with hard business rules (eligibility, credit policy) and diversity quotas across product types via constrained re-ranking.
   • Show how to implement a two-stage scorer + constrained re-ranker.
6. Online Inference Architecture (≤100 ms p95)
   • Outline the serving architecture including feature store, candidate generation, and fallbacks when services degrade.
7. Evaluation
   • Specify offline metrics (e.g., NDCG@k, MAP, calibration) and online guardrails (latency, application error rate, CS contacts).
8. Monitoring and Rollback
   • Detail monitoring for model/data drift and a rollback plan.