•Designed and automated ETL/ELT data pipelines (Python, SQL) to ingest and transform 20k+ financial data sources from REST
APIs (e.g., FRED, Morningstar) into Oracle/PostgreSQL Autonomous Datawarehouse, achieving 99.9% reliability.
•Migrated MongoDB to a centralized data warehouse on Oracle Cloud and built cross-database synchronization for seamless data
integration, implemented data quality checks, scheduled daily updates, incident alerts (Slack, E-mail), and failure retry logic,
reducing product error rate by 90%.
•Engineered CSORE, a profitability composite indicator synthesizing peer group metrics and financial indicators reflecting company
profitability and quarterly performance to improve EPS (Earnings Per Share) prediction accuracy.
•Applied NLP (prompt engineering, few-shot learning with GPT-4o, FinBERT) to extract evasiveness/bullishness/sentiment from
earnings call transcripts, enhanced EPS predictive modeling and integrated results into warehouse tables for downstream BI.