Amritanshu

Mumbai

Summary

Data Analyst with 1.5 years + experience at Kotak Mahindra Bank, where I worked with corporate salary account data across 20M+ customers to help teams make faster, smarter decisions. I built dashboards that branch managers on pan india level use daily, cut a reporting process from a full day to 30 minutes, and led a large-scale data migration to AWS. I work mostly in SQL and Python, and I'm comfortable taking a messy business problem and turning it into something actionable.

Overview

year of professional experience

Certification

Work History

Data Analyst

Kotak Mahindra Bank Ltd.

Mumbai

12.2024 - Current

Analyzed bonus credit patterns across 2.5M+ corporate salary accounts to identify the best windows for customer targeting . Directly influenced campaign timing strategy and deposit growth.
Built and owned a full end-to-end Power BI dashboard (connected to Azure + SQL, refreshing daily) to track customer credit inflows and outflows.The dashboard became the go-to tool for Relationship and Branch Managers to proactively spot at-risk accounts.
Delivered monthly reports on customer acquisition trends to senior stakeholders, translating data signals into actionable portfolio strategy recommendations.
Led the migration of 5,000+ lines of SAS/PROC SQL code to MySQL on AWS Redshift, maintaining data integrity throughout a business-critical platform transition and improving query performance in the process

Education

B.Tech - Electronics And Communications Engineering

Birla Institute of Technology

Mesra

06-2023

Skills

BI & Reporting: Power BI, MS Excel, Tableau, Stakeholder Reporting
Languages & Querying: Advanced SQL, Python (Pandas, NumPy)

Machine learning (Intermediate Level)
Cloud & Data Ops: Microsoft Azure, Apache Airflow, AWS (Redshift, S3), SAS

Projects

Deposit Book Growth Dashboard

Built a automated daily-refresh Power BI dashboard that flagged customers with overnight large credit inflows like (salary, bonus, or lump-sum) as high-potential FD/RD conversion targets.
Gave Branch Managers a prioritized call list every morning, directly supporting the bank's liability growth strategy by connecting data signals to frontline sales actions.

Customer Churn Prediction

Framed churn as a business problem first: identified that missing high-risk customers was costlier than false positives, then built an XG-Boost model optimized for recall (80%) — correctly flagging 530 of 670 actual churners on a 10,000 record dataset.