At Think360.ai, I work with banking transaction data for ML training and downstream workflows, using SQL to extract, validate, and investigate records. I also validate REST API responses against PostgreSQL data and perform reconciliation and data-quality checks.
For the SBI YONO 2.0 PFM feature, I contributed to banking transaction classification work associated with a reported 20% improvement in model accuracy. I also contributed to a FastText validation layer alongside an SVM classifier that routed prediction disagreements for human review, reducing human intervention by approximately 85%; outside work, I'm developing a bank-statement analysis application.
