EY·Data Scientist·Technical Phone Screen·Senior

Senior

May 2026

Summary

EY data scientist interview that threw a pretty gnarly domain knowledge question at me. One question, heavily cross-domain, and it required knowing actual financial services internals rather than just generic data modeling concepts.

Questions Asked (1)

Pick two financial services domains (from Consumer, Commercial, Capital Markets, Risk & Compliance, Finance, Enterprise, Life Insurance, P&C Insurance, Wealth, or Regulatory Reporting) and compare them across three dimensions: the core data model entities and grain, the typical control points and reconciliations, and the most common failure modes. Then propose a reusable pattern or schema that spans both domains and explain where it breaks down.

Data ModelingSystem DesignTechnical Trade-offs

Author's notes

This was a lot.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Choose two domains you know well (e.g., Consumer and Capital Markets) and structure your answer by first defining each domain's data model, controls, and failure modes, then synthesizing a cross-domain pattern. Emphasize the trade-offs and where the pattern breaks, showing you understand both technical and business nuances.

Pro tip: Anchor your comparison in real-world examples from your experience, and explicitly discuss how regulatory requirements (e.g., BCBS 239) shape data models and controls differently across domains.

1. Select and Justify Domains

Pick two domains that contrast well (e.g., Consumer vs. Capital Markets) and briefly explain why they are relevant to the role and interesting to compare.

2. Compare Data Models

For each domain, describe the core entities (e.g., Customer, Account, Trade, Position) and the grain (e.g., daily snapshot, transaction-level), highlighting differences in granularity and relationships.

3. Compare Controls and Reconciliations

Outline typical control points (e.g., data validation, reconciliation, audit trails) and reconciliation processes (e.g., intraday vs. end-of-day) for each domain, noting regulatory drivers.

4. Compare Failure Modes

Identify common failure modes (e.g., data quality issues, latency, mismatched trades) for each domain and explain how they manifest differently.

5. Propose Cross-Domain Pattern and Limitations

Suggest a reusable pattern (e.g., a canonical data model with domain-specific extensions) and discuss where it breaks down due to domain-specific constraints or regulatory requirements.

Key Points to Mention

Core entities and grain differences (e.g., Consumer: Customer/Account at daily grain; Capital Markets: Trade/Position at transaction grain)
Control points such as data validation, reconciliation, and audit trails, and how they vary by domain
Failure modes like data quality issues, latency, and reconciliation breaks, with domain-specific examples
Regulatory influences (e.g., BCBS 239, GDPR, MiFID) on data models and controls
Reusable pattern ideas: canonical data model, data vault, or domain-driven design with bounded contexts
Where the pattern breaks: conflicting grain requirements, regulatory silos, or real-time vs. batch processing needs

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.