Shubham Kumar

I’m a data analytics professional who brings hands-on experience in SQL, Python, and Power BI reporting, with exposure to ETL pipelines, data validation, and machine learning. I’ve built ETL pipelines, star-schema models, and classification models using real-world datasets.

At Outlier AI, I worked as an AI Language Model Trainer, performing data validation and quality checks on hundreds of structured prompt-response datasets weekly. I applied structured evaluation criteria and scoring rubrics to identify factual errors and logical inconsistencies, supporting consistent feedback workflows for AI output review.

In my clinical experience at Gastrolab Pvt Ltd, I performed data validation and integrity checks on AI-assisted EGG diagnostic outputs, ensuring dataset accuracy by comparing discrepancies against patient history records. I also analyzed structured clinical datasets to surface diagnostic patterns and supported medical-team review with cleaned, structured reporting outputs.

I further strengthen my technical foundation through projects like an end-to-end Blinkit Retail Analytics Pipeline (ETL, star schema modeling, and automated KPI reporting in Power BI) and Multi-Model Fraud Detection for 100K+ transactions using ROC-AUC and threshold optimization. I’m also a first-author on a peer-reviewed ACS Omega review article covering computational and AI-driven approaches in nanomedicine.