Kumar Tholeti

Data Scientist @ANZ

Led a data science team within the Asset and Wealth Management division, overseeing the delivery of AI/ML solutions. Established an end-to-end data processing pipeline, encompassing data collection, cleaning, and preprocessing. Fine-tuned pre-existing machine learning models using transfer learning techniques to address business needs Discovered relevant patterns within the data to inform optimal decision-making aligned with business objectives. Selected appropriate machine learning models to address specific business challenges. Trained various supervised and unsupervised machine learning models. Introduced a document data extraction engine using advanced NLP techniques, reducing manual document review time by 70%. Developed an automated ID document verification engine based on Computer Vision machine learning, saving 2 million USD/year, and reducing reliance on third-party tools. Successfully tackled natural language processing (NLP) problems using advanced machine learning models like transformers, achieving state-of-the-art results. Fine-tuned pre-existing machine learning models using transfer learning techniques to address business needs. Deployed machine learning models on cloud platforms, including AWS. Established an end-to-end machine learning ecosystem for client onboarding, resulting in a 60% reduction in onboarding time and improved customer waiting times. Developed and implemented machine learning models for business use cases using Python, NumPy, Pandas, SciPy, Scikit-learn, TensorFlow, Keras, Pytorch, and other libraries as necessary. Implemented a QA-based OCR engine for high-quality data extraction. Successfully created an automated payments engine handling 7k payments per day. Decreased client attrition by 50% through the analysis of client behavior patterns.