Remote Data Analyst - AI Evaluation (Cloud Data Warehouse Focus) at Henry Hire
JOB DESCRIPTION
micro1 is staffing a client project that stress tests AI assistants against the kind of analytical work real data teams do every day: pulling from cloud data warehouses, investigating anomalies, building KPI reports, and validating that AI-generated answers actually match the underlying data.
You won't be building AI. You'll be the domain expert who catches what it gets wrong. No AI background needed--what matters is that you know your way around SQL, Snowflake, and real analytics workflows well enough to spot a bad join or a mistimed aggregation on sight.
What You'll Actually Do
Run realistic analytical scenarios Anomaly investigation, KPI reporting, and data refreshes using AI tools connected to live cloud data warehouses.
Fact-check the AI. Write your own SQL against source data to verify every figure it produces, and scrutinize its joins, filters, and time windows for subtle errors.
Own the test environment. Maintain, reset, and manage seeded datasets in Snowflake so each test scenario starts from a clean, correct state.
Manage access. Set up and oversee warehouse roles, permissions, and access controls to keep evaluation environments secure and repeatable.
Handle the plumbing. Configure and document authentication and connectivity across multiple analytics and SaaS platforms.
Flag the weird stuff. Document any undocumented or unexpected product behavior you run into along the way.
Calibrate with the team. Join peer sessions to keep grading and scoring consistent across evaluators.
What Gets You Hired
3+ years as a data analyst or analytics engineer with advanced SQL and real Snowflake experience--warehouses, access control, query history, not just SELECT statements.
A sharp eye for reconciliation. You've audited reported metrics against raw data before and know exactly where aggregation logic tends to break.
Business/finance analytics fluency. You understand how KPIs are defined and reported to leadership or external stakeholders, not just how to query them.
Database administration chops. Managing roles, grants, and authentication protocols. Bonus points for OAuth or security-integration experience.
Comfort with everyday SaaS tools. Slack, Google Workspace, or Microsoft 365 for sharing findings.
Hands-on AI assistant experience for analytics. Ideally with a healthy skepticism toward AI-generated SQL and a track record of catching its mistakes.
QA or rubric-based evaluation background. Data labeling, structured grading, or similar with meticulous attention to detail and strong written and verbal communication.
RESUME TIPS
How to tailor your resume for this role at Henry Hire.
Lead with SQL specifics, not just "SQL skills." Instead of "proficient in SQL," write "wrote complex SQL queries to reconcile KPI dashboards against raw warehouse data" or "built and maintained data validation queries in Snowflake." Specificity signals real hands-on work.
Name Snowflake explicitly if you've used it. If your warehouse experience is in Snowflake, say so by name--don't bury it under generic "cloud data warehouse experience." If your experience is in BigQuery, Redshift, or similar, mention it anyway and note the overlap in concepts (warehouses, roles, query history).
Quantify a time you caught an error. A line like "identified a mismatched join that was overstating monthly revenue by 8%" is far more compelling than "strong attention to detail." This role is fundamentally about catching mistakes--show, don't tell.
Call out access/role administration if you've done it. Even limited experience managing warehouse roles, grants, or permissions is worth its own bullet point--many analysts never touch this side of the stack, so it differentiates you.
Mention any AI tool usage AND your skepticism of it. If you've used Copilot, ChatGPT, or an AI analytics assistant and caught it making an error, say so directly: "used AI-assisted SQL generation, identified and corrected incorrect date-window logic in three instances."
Surface QA, audit, or labeling experience even if it's not your main title. If you've done rubric-based grading, structured QA, or data labeling as a side responsibility, pull it into its own line. It's a specific qualification this role is filtering for.
Translate KPI experience into stakeholder language. Instead of "built reports," try "delivered KPI reporting reviewed by leadership" or "owned metrics reporting for external stakeholders" as this signals you understand the business side, not just the query side.
List specific SaaS tools by name. "Slack, Google Workspace" is a quick, easy match for an ATS scan and a human reviewer. Don't just say "collaboration tools."
Keep it to one page if possible, and cut anything that isn't analytics, SQL, warehousing, QA, or AI-adjacent. This is a specialized evaluator role. A resume cluttered with unrelated experience dilutes the signal.
Last but not least: THERE WILL BE AN ASSESSMENT TEST. They will detect if you wrote SQL queries using AI, and it will fail you. If you must use AI to help write queries for the exam, then modify the AI output then run it through an AI detector like 'MyDetector' to ensure you modified it enough to pass as human written.