Shahrukh Khan Email & Phone Number

Bengaluru, Ka, In

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

"Financial data pipeline for a leading financial data aggregation and analytics platform (20+ year partnerships with top US banks) using NLP & ML based Transaction Data Enrichment model."Implemented unsupervised ML based merchant name grouping pipeline, standardizing ~99k variations (2.4%) in a 4.1M long merchant name dataset and optimizing it for reduction of execution time by 98% (25 days to 14 Hours).Built an XGBoost classifier for imbalanced data. Employing hyperparameter tuning, feature engineering and SMOTEENN (imbalanced-learn for combined data oversampling and undersampling), improving the f1-score by ~18%.POC of an in-house financial text translation solution using Hugging Face Transformers & NLP achieving a 100% cost reduction and 80% time reduction compared to third-party APIs.Identified and resolved critical misclassification issues in the enrichment pipeline (keyword analysis, debugging) and is_physical module, escalating findings and solutions to the principal data scientist.

Bengaluru, Karnataka, India

Built a cost-effective web scraping module to iteratively extract latitude and longitude data for 38,000 US cities from public sources, enriching our data pipeline with valuable geospatial information.Optimized model performance (accuracy, recall, fill rate) through comparative analysis and collaboration with module owners. Additionally, automated model validation pipeline for geographically specific models. Supported online/offline merchant prediction through data signal analysis, presenting findings & recommendations to principal data scientist.