Axiom Command Deck

Competencies

weights sum to 1.00

Name	ID	Weight	Definition
Distributed Systems & Data Platform Architecture	distributed_systems_architecture	0.30	Design and own high-throughput data pipelines (10TB+/day scale) with schema evolution, partitioning strategies, and lineage tracking. Make principled decisions on tool selection (Kafka, Flink, Iceberg) and justify cost/latency/reliability tradeoffs for 800+ enterprise customers.
SQL Performance & Query Optimization	sql_query_optimization	0.20	Read EXPLAIN plans, diagnose inefficient joins and aggregations, and rewrite queries for production workloads. Optimize warehouse costs by identifying and fixing expensive queries across analytics, ML features, and billing pipelines.
Technical Leadership & Mentorship	technical_leadership	0.25	Guide 4 mid-level engineers through system design decisions; review complex architecture docs and production handoffs. Lift team velocity by establishing data contracts, backfill strategies, and operational patterns that prevent costly incidents.
Streaming & Batch Processing Systems	streaming_batch_systems	0.15	Hands-on expertise with one or more stream processors (Flink, Spark, Beam) to build fault-tolerant, low-latency ingestion. Handle backfills and schema migrations without downtime in a production warehouse serving live analytics and billing.
Data Infrastructure Cost Optimization	cost_optimization	0.10	Identify and execute initiatives that reduce six-figure monthly warehouse spend through smarter partitioning, storage formats (Iceberg), and query patterns. Model infrastructure costs and tradeoffs between compute, storage, and latency.

Scoring Weights

Adjust how much each competency contributes to the final score. Changes apply to all future evaluations for this job spec.

distributed_systems_architecture

30%0.30

Scoring guideline

sql_query_optimization

20%0.20

Scoring guideline

technical_leadership

25%0.25

Scoring guideline

streaming_batch_systems

15%0.15

Scoring guideline

cost_optimization

10%0.10

Scoring guideline

Total weight: 1.00 ✓ balanced

Interview questions

AI-generated from the JD · auto-published · edit anytime

Question count3 currently · candidate sees this many

Title

Difficulty

Time budget (min)

Prompt

You're migrating a user events table from a flat schema to a nested structure. The old schema has columns: event_id, user_id, event_type, property_key, property_value (one row per key-value pair). The new schema uses a MAP<STRING, STRING> column: event_id, user_id, event_type, properties. You must backfill 2 years of historical data (500M rows) while live events continue writing to both schemas via dual-write. Write a SQL query that: (1) groups old events by event_id and aggregates properties into a single MAP, (2) ensures no data loss if a backfill job restarts mid-execution, (3) handles late-arriving events that arrive during backfill. Assume a batch_id column marks backfill chunks. Example: Old rows {event_id=1, user_id=100, event_type='click', property_key='url', property_value='example.com'} and {event_id=1, user_id=100, event_type='click', property_key='button', property_value='submit'} should produce one row {event_id=1, user_id=100, event_type='click', properties={url→example.com, button→submit}}.

Starter code

Hidden tests

Hints (one per line)

Behavioral round (Round 02)

Voice Q&A · pre-generated from the JD · edit anytime

4 prompts · Cal walks through each in order during Round 02

Prompt

Target competency

Time (min)

Follow-up (optional)

Live performance

Aggregated from candidates who took this job spec.

Interviews

1

0 completed

Avg score

—

out of 100

Completion

0%

0 of 1

Avg score per competency

Distributed Systems & Data Platform Architectureweight 0.300.17n=6

SQL Performance & Query Optimizationweight 0.200.17n=6

Must-haves

required8+ years in data engineering or distributed systems
requiredProduction experience designing for schema evolution and backfills without downtime
requiredHands-on expertise with at least one streaming or batch processing framework (Spark, Flink, Beam, or dbt-at-scale)
requiredDemonstrated track record of mentoring engineers and improving team output

Staff Data Engineer

Competencies

Scoring Weights

Interview questions

Behavioral round (Round 02)

Live performance

Must-haves