NIKUNJ MAHESHWARI

Managed and optimized critical data pipelines for Target's finance team, encompassing Fixed Assets, Vendor Income, and Department Profit & Loss portfolios, ensuring data integrity and financial reporting accuracy.

Designed and developed scalable, high-performance data pipelines using Apache Spark and Hadoop, significantly enhancing processing capabilities for large datasets.

Optimized Spark job performance by tuning configurations (memory, shuffling, partitioning), leading to improved cluster resource utilization and efficiency.

Implemented robust Data Quality (DQ) checks and automated alert systems via Slack and email, enabling rapid discrepancy resolution.

Developed reusable Spark functions for common ETL operations, resulting in significant development time savings for future data projects.

Managed the end-to-end back-end risk infrastructure for Udaan Capital's digital lending products, overseeing INR 3 billion AUM and ensuring system stability and compliance.

Led development of critical risk infrastructure components, including the customer onboarding engine, line state management, and dynamic limit revision systems.

Engineered an automated end-to-end limit revision engine, integrating diverse alternative data sources (transaction, GSTIN, bank statement APIs, SMS data) to dynamically adjust customer credit limits.

Designed and coded a real-time onboarding algorithm for Udaan's lending marketplace, enabling real-time underwriting and reducing loan disbursal Turnaround Time (TAT) by 75%.

Deployed the Risk Engine as a Flask-based microservice, enabling real-time credit application assessment for seamless lending operations.

Designed and deployed an XGBoost model with 80% accuracy, leveraging 500+ features to predict buyers exceeding 30+ Days Past Due (DPD).

Implemented an early-intervention contact strategy for collections agents, reducing DPD outstanding amounts by 20%.

Built an extensive feature space that served as the primary data source for analytics, data science, and reporting teams.

Designed and implemented scalable Apache Spark (Scala) data pipelines for over 350 million customers, enabling advanced feature engineering, modeling, and customer segmentation.

Optimized data pipelines to achieve near real-time data processing for Paytm Postpaid customers, enhancing operational efficiency.

Developed an automated hourly ETL script to aggregate Postpaid data and push it to Elastic Search clusters, improving efficiency and enabling real-time analytics.

Delivered multiple critical data warehousing and analytics projects in collaboration with cross-functional stakeholders, gaining deep expertise in Hadoop Architecture, including storage management and optimization.

Promoted ahead of schedule and awarded the 'Disruptor and Learner Award' in recognition of significant contributions across multiple initiatives within the cohort.

Analyzed customer issues across various domains (product, orders, deals, RMA) to generate actionable business insights for the Customer Care Centre Business Layer.

Led the end-to-end release of a strategic project, ensuring timely delivery under tight deadlines.

Designed and implemented robust ETL pipelines to migrate data from C3 OLTP systems to Hadoop for analytical processing.