- Spearheaded PayU's Cash-Loans and BNPL data strategy and decision-making, closely collaborating with cross-functional teams to drive growth, risk, core product, engineering, business, and operations.- Developed a lending intelligence framework that enhances all user lifecycle stages from acquisition to retention, including disbursals, collections, propensity, cross-selling, and targeting underserved segments to expand our consumer base.- Collaborated with Legal, Compliance, and Privacy teams to establish best practices and lead initiatives in data governance and MLOps frameworks and platforms, streamlining model deployment, monitoring, and change management.- Designed a real-time decision engine using transformer-based information extraction techniques, leveraging diverse data sources such as bureau, SMS, geo, app tracking, banking, transactions, and payment data to optimize lending decisions.- Built Customer Lifetime Value framework for better acquisition of BNPL Customers which optimises retention, spends, cross-sell, and risk.- Implemented a robust Fraud Framework which had components like fraud score, anomaly & forgery detection, vulnerability testing, and dedupe solution using SOTA technologies of vector databases and LLM based embeddings, achieving an 85% reduction in identity fraud.- Led and mentored a diverse team of Data Scientists, ML Engineers, and Product Managers, fostering a data-aware culture that enhances collaboration, boosts data capabilities, and promotes data literacy across the organization.

In association with Great Learning

• Responsible for building location-based intelligence/ Geo-Spatial Capability across BFL including customer hotspot identification, address change prediction, building in-house dashboards using OSM data, smart address creation etc.• Fraud detections for customer location using Density-based spatial clustering of applications with noise(DBSCAN).

- Built a high-performance Data Feature Lake achieving 3600 unified views across customer, network, and device data, harnessing IoT data streams and predictive analytics to mitigate churn and enhance customer service.- Engineered scalable, custom ML algorithms using Map-Reduce and distributed programming on one of the largest Hadoop clusters for 400+ million customers, optimizing real-time data integration and behavior modeling.- Implemented Campaign Analytics utilizing time series modeling and deep learning for precision in inventory optimisation and demand forecasting.- Designed Anomaly Detection services to monitor customer/ network behaviors, to address security & operational anomalies.- Built Network Intelligence using customer geo-data integrated with IoT sensors to optimize tower placements and bandwidth allocation to facilitate proactive network management and failure prevention, showcasing a strong alignment with Industry 4.0 principles of interconnected, automated, and optimized infrastructure management.

- Created language models from text data(job descriptions) which draws heavily from probabilistic (Naive Bayes), Sequential (LSTM) as well as rule-based models to build resume classifiers.- Worked on Text analytics including the processing of text, tokenization, word2vec, parsing, parts-of-speech (POS) tagging, named entity recognizer (NER), chunking, sentiment analysis, topic modeling, etc. - Developed an Annotation tool for tagging text corpus to help Information retrieval and text mining.- Automated the recruitment process using Talent data from systems like CRM, ATS, etc.

• Actively involved in the full development cycle for ETL, end to end from design, implementation, Validation and testing, documentation, and maintenance. • Responsible for mining & analyzing large data by performing Data Integration, Cleaning, Pre-processing, Feature Transformation, Hypothesis testing, Model Creation, Evaluation and performance optimization.• Founded ML club to re-skill freshers in niche technologies like Data Science, Machine Learning, IoT etc.• Developing hypotheses and proving them using conventional Machine Learning and Neural Network models, Regression Models, k-NN.

Research and development in the conventional methods of Image Processing. Face Recognition and Medical Image Analysis were the major areas of focus.