← Excel Impact Interview Insights

Excel Impact·Technical Product Manager·Technical Phone Screen·Senior

Senior
May 2026

Summary

Interviewed for a TPM role at Excel Impact. One technical question about ETL tooling stood out, felt like they were probing for real-world troubleshooting instincts rather than textbook answers.

Questions Asked (1)

Q1

AWS Glue isn't cutting it for your workload. What alternative ETL tools would you look at for handling unstructured data in a data warehouse context?

Technical Trade-offsSystem DesignRoot Cause Analysis
Author's notes

I rambled a bit here.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the specific limitations of AWS Glue for unstructured data (e.g., schema enforcement, performance, cost). Then propose alternative ETL tools that better handle unstructured data, such as Apache Spark, Databricks, or AWS Glue with custom scripts, and discuss trade-offs in terms of scalability, cost, and integration with the data warehouse. Emphasize a product manager's perspective: aligning tool choice with business needs, team skills, and long-term maintainability.

Pro tip: Frame your answer around the 'build vs. buy' decision and total cost of ownership, not just technical features. Show you understand that the best tool depends on the team's expertise, existing infrastructure, and the specific nature of the unstructured data (e.g., text, images, logs).

1. Clarify the problem

Ask clarifying questions about the workload: what types of unstructured data, volume, velocity, and the target data warehouse. This shows you don't jump to solutions without understanding requirements.

2. Identify Glue's limitations

Briefly state why Glue isn't cutting it: e.g., limited support for complex transformations on unstructured data, performance bottlenecks, or cost at scale. This justifies the need for alternatives.

3. Propose alternatives

Suggest 2-3 alternative ETL tools (e.g., Apache Spark with custom code, Databricks, Apache NiFi, or cloud-native services like AWS Lambda with S3 triggers). Explain how each handles unstructured data better.

4. Evaluate trade-offs

Compare alternatives on dimensions like scalability, cost, ease of use, integration with the data warehouse, and team expertise. Highlight that the choice depends on specific constraints.

5. Recommend and justify

Make a recommendation based on the context, and explain how you would validate it (e.g., proof of concept, cost analysis). Show product thinking by considering stakeholder needs and long-term strategy.

Key Points to Mention

  • Apache Spark (with Python/Scala) for flexible processing of unstructured data
  • Databricks for managed Spark with collaborative notebooks and optimized performance
  • Apache NiFi for data flow automation and handling diverse data formats
  • AWS Glue with custom scripts or Glue Studio for semi-structured data
  • Consideration of data warehouse integration (e.g., Redshift, Snowflake) and ETL vs. ELT patterns
  • Total cost of ownership, including development, maintenance, and operational overhead

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