Responsibilities:
Write advanced SQL queries (using JOINs, CTEs, WINDOW functions, and subqueries) to pull precise training slices from the data warehouse.
Use SQL string functions (SPLIT) to flatten nested JSON blobs, logs, or free-text fields into tabular rows ready for labeling.
Apply CASE statements and conditional logic to normalize dirty data before they reach the annotation queue.
Manually review queried data and assign granular labels.
Write dynamic SQL rules to automatically pre-label high-confidence data so human annotators only review the ambiguous.
Build automated SQL audit queries that can scan the annotated dataset daily to flag duplicates, conflicting labels, null values, or out-of-distribution samples before they reach the ML Team.
Partner with ML Engineers to test proposed annotation guidelines against actual database records.
Requirements:
Bachelor’s degree in Computer Science or a related field.
5+ years’ experience writing advanced SQL in a professional setting and can be able to identify the difference between INNER JOIN, LEFT JOIN, and LATERAL FLATTEN without Google.
Experience with annotation tools (Labelbox, Super Annotate) and familiar with Python or R for Data manipulations.
Ability to spot a missing comma or a mislabeled data point from a mile away and understand the ripple effect it has on model performance.
Ability to translate technical database structures into plain English annotation guidelines for non-technical team members.
Renumeration : NGN 500,000

Company in Nigeria.