How to assess · For hiring teams

How to Assess Snowflake / BigQuery Skills When Hiring

The test formats that actually work for Snowflake / BigQuery, what a strong answer looks like, sample questions and a scoring rubric you can use as-is.

The short answer

Assess Snowflake / BigQuery with a task, not a conversation: diagnose an expensive query, design a data layout, ai-scored assessment (e.g. cohesyve) or cost investigation. Score it against written criteria you fix before you see any submissions, and weight the criteria that the role actually depends on.

Understands the pricing model — compute, storage, scanned bytes — and designs with it in mind
Lays out data with partitioning and clustering that match query patterns
Reads a query profile and can say why a query is slow or expensive
Sizes and schedules compute appropriately, and turns it off

Build a Snowflake / BigQuery assessment →Read the full article ↓

Paste a job description; Cohesyve generates a role-specific assessment and rubric. Ten candidates free, no card.

Cloud warehouses changed the economics of analytics and created a new way to fail: bills. A query that scans a terabyte to return ten rows, a warehouse left running, a clustering key chosen badly — each is a cost that arrives at month end. The skill for Snowflake, BigQuery and their peers is not SQL; it is knowing how the engine charges, how data should be laid out, and how to keep a platform fast and affordable as usage grows. This page covers how to assess cloud data warehouse skill for data engineering, analytics engineering and platform roles.

Why Snowflake / BigQuery is worth testing

Warehouse mistakes are invisible until the invoice. A team that does not understand pruning, partitioning and compute sizing can spend ten times what it should while wondering why queries are slow. Testing shows whether a candidate reasons about cost and performance together, and that determines whether the platform scales or gets throttled by finance.

What strong Snowflake / BigQuery looks like

Understands the pricing model — compute, storage, scanned bytes — and designs with it in mind
Lays out data with partitioning and clustering that match query patterns
Reads a query profile and can say why a query is slow or expensive
Sizes and schedules compute appropriately, and turns it off
Manages access with roles and least privilege
Uses warehouse-native features — semi-structured data, time travel, materialised views — with judgement
Monitors cost and performance and can attribute spend to teams

Ways to assess Snowflake / BigQuery

Diagnose an expensive query

Provide a query, its profile and the table layout for a query that scans far more than it needs. Ask why and what to change — query, layout, or both.

Pros

The core skill; scoreable.

Cons

Engine-specific fixture.

Best for Any level.

Design a data layout

Describe a large event table and its common queries. Ask for partitioning, clustering and any pre-aggregation, with reasons.

Pros

Tests design judgement.

Cons

Design-based.

Best for Mid and senior roles.

AI-scored assessment (e.g. Cohesyve)

Generate a cloud data warehouse task from the job description — a cost diagnosis, a layout design, an access-control question — with a rubric. Each candidate receives a different variant; the reasoning is scored alongside the work.

Pros

Asynchronous and consistent across a large pool; a different task per candidate removes shared answers; scores the explanation, which is where judgement shows.

Cons

Cannot run tools on the candidate's behalf; keep a human review for shortlisted finalists.

Best for Screening an applicant pool fairly before interview time is spent.

Cost investigation

Present a monthly bill breakdown that doubled and ask how they would find the cause.

Pros

Tests operational maturity.

Cons

Needs realistic data.

Best for Platform owners.

Cohesyve

Run a Snowflake / BigQuery assessment on your next opening

Cohesyve generates a unique Snowflake / BigQuery task per candidate from your job description, with the scoring rubric attached. Questions are different for every applicant, so they cannot be shared or looked up.

Try it free →Book a 30-min demo

What to test

Cost model

Whether they know what they are paying for.

Explain what drives cost in the engine you useEstimate the cost of a given queryReduce spend without slowing users

Data layout

Whether data is arranged for the queries.

Choose a partition key for an event tableDecide when clustering helpsExplain pruning

Query performance

Whether they can fix slow queries.

Read a profile and find the expensive stepRewrite a query to pruneDecide when to materialise

Operations and governance

Whether the platform is managed.

Set up role-based accessAttribute cost to teamsSchedule and size compute

Sample Snowflake / BigQuery questions

What determines the cost of a query in your warehouse?

Entry

Look for Bytes scanned or compute time depending on engine; storage separately; design implications.

A query on a date range scans the whole table. Why, and what do you change?

Mid

Look for No partitioning on date, or the filter prevents pruning; partition and write prunable predicates.

When does clustering help, and when is it wasted?

Mid

Look for High-cardinality columns filtered often on large tables; wasted on small tables or unfiltered columns.

The monthly bill doubled. How do you find out why?

Mid

Look for Query history by cost and user, new workloads, runaway schedules, compute left running, storage growth.

Design access so analysts can read marts but not raw PII.

Senior

Look for Roles per layer, column masking or views, least privilege, audit.

Red flags

Cannot explain what a query costs
Selects star from large tables habitually
Has never read a query profile
Leaves compute running
Grants broad access for convenience

Scoring rubric

Criterion	Weight	What strong looks like
Cost awareness	30%	Designs and queries with the bill in mind.
Data layout	25%	Partitioning and clustering match access patterns.
Query diagnosis	25%	Reads profiles and fixes causes.
Operations	20%	Access, sizing and monitoring are managed.

Mistakes hiring teams make

Testing SQL rather than the engine's model
Ignoring cost entirely
Not including a layout design
Accepting a correct query that scans everything
Confusing certification with platform judgement

Roles that need Snowflake / BigQuery

Data EngineerAnalytics EngineerData Platform EngineerBI DeveloperData ArchitectCloud Engineer

Common questions

Snowflake or BigQuery — does the assessment differ?

The pricing models differ and so do some layout features, but the judgement — prune, lay out for queries, manage compute and access — is the same. Test the engine you run.

What is the best single warehouse question?

Show a date-range query that scans the whole table and ask why. Pruning understanding is the foundation of both performance and cost.

How much warehouse knowledge should an analyst have?

Enough to write prunable queries and know that a select star on a big table costs money. Layout and governance are platform skills.

How long should a warehouse assessment take?

Forty-five minutes for a diagnosis-and-layout exercise.

Cohesyve · Skill assessments for hiring

Test Snowflake / BigQuery before the first interview

Generate a role-specific Snowflake / BigQuery assessment from your job description and see who can do the work before you spend interview time on them.

1,500+

assessments completed

50%

faster time-to-hire

90%

completion rate

5 min

from JD to assessment

No credit card · 10 free candidates · Plans sized to your hiring volume

From the blog

Related skill guides