• Built time series regression models for annual financial planning; forecasted checking fees portfolio worth more than $500M/year• Incorporated various drivers including enrollment, transactional, business and finance attributes• Successfully achieved 84% accuracy for overdraft income models; overdraft income had been highly unstable with after Covid• Worked as the main point of contact for leadership, external teams like MRM, finance; successfully presented outcomes to extended senior leadership• Transformed checking income portfolio reporting using PowerBI, Python and SQL; established real time portfolio update incorporating finance and transactional data sources• Automated portfolio income/expense distribution process to seamlessly integrate with multiple teams and their platform dependencies• Proposed new product features analyzing large datasets containing 2B+ records; resulted in $200M.year in potential member savings• Provided end to end analytics support for various marketing campaigns for new products/offers; $15B+ new money acquired from these campaigns• Developed financial distress model identifying members with increased risk of overdraft fees; successfully eliminated 18% of overdraft fees• Led the efforts to redefine and target engaged members; engagement share increased by 7% and 1M new members chose NFCU as their preferred financial institution

• Analyzed returned item fee data to identify repeated payment attempts from vendors; saved $50M+ in repeated fee attempts• Used various fraud detection modeling techniques to identify accounts and patterns of checking account fraud; discovered $200M+ and 20K+ fraudulent accounts engaged in fraudulent transactional activities• Supervised efforts to revamp deposit portfolio end to end reporting; optimized data architecture and transformation logic to increase real time availability to 300%• Created a user level insight data lake with predicted features; insight measures to be used in 10+ campaigns by various teams • Presented on various internal conferences and performed duties of judge and coach for hackathons

• Developed data pipelines for analytics platform; transformed Excel reports into interactive dashboards• Performed end to end Zendesk integration process with University website; created interactive dashboards using complex SQL queries on Zendesk Explore analytics • Built a self service customer support tool to reduce workload on employees; reduced employee efforts on student concerns by 40%• Utilized information articles to build a simple text based recommendation model to feed into self serve customer service bot; successfully solved 24% of questions submitted on support bot• Managed 40000 campus residents and housing committees’ data; generated weekly Power BI reports • Architected relational data models for 9 on-campus residence halls; developed custom APIs to facilitate seamless data transfer across multiple platforms • Combined historical unstructured employee data from various legacy system; established data pipelines from SQL server to Quickbase application • Arranged online training sessions; trained employees on PowerBI and SQL queries

• Developed data pipelines for analytics platform; transformed Excel reports into interactive dashboards• Built machine learning models using XGBoost, Random forest and Kernel SVM algorithms for cost prediction; achieved 0.23 root mean square error • Designed custom APIs in Spotfire application linking SQL data with Python scripts; utilized Spotfire python interpreter to reduce process time by 30% • Analyzed 3rd party vendor data for risk identification; reported high risk vendors using PowerBI Dax queries• Created data architecture to seamlessly feed public and vendor data into secure analytics space; constructed data lake involving raw data elements and predicted feature variables to be used for reporting• Designed self service analytics platform where users can upload their oil and gas projects and the platform will predict the total cost including all individual component costs• Built a functionality in the analytics platform where users can compare their costs with their competitor projects on a component level as well as whole project• Integrated an open source python interpreter inside the platform; developers can use existing data points to write models and save the output variable as a separate data point for their projects• Led a team of 5 interns; analyzed and reported high risk 3rd party vendors using PowerBI Dax queries

• Cleaned raw transaction data using SQL queries, visualized hourly transaction summary on Tableau• Engineered Hyperledger architecture to enhance system concurrency and performance; number of simultaneous stakeholders increased by 40% • Developed NOSQL database architecture with CouchDB; enhanced transactions success by 10 transactions/ second• Liaised with client, project manager and backend team; Oversaw 50 hours/week implementation effort using Jira• Designed respossive wireframes/prototype using Invision and Balsamiq; lowered front end development efforts by 30% and HTML efforts by 40%

• Developed Employee salary management system in Python, utilized Pandas and NumPy libraries for rapid and accurate computations; manual efforts of HR work reduced by 2 hours per day• Redesigned system using Django framework; user traffic and response rate augmented to 10% due to concise and thorough user experience