top of page

Sanchit Suman

Professional Experience

2020 - Present

Data Scientist, Apple - AMP financial fraud

  • Key contributor in design and development of an analytical credit card dispute management system that now saves over $30 million every year through targeted challenges of credit card disputes. 

  • Designed and implemented dispute management performance monitoring to identify long-term trends and opportunities using Python, SQL and Tableau. This was also extended to identify aberrant patterns whose resolution led over savings of $2M/year. 

  • Designed and implemented analytical process to calculate KPIs that drove influential conversations with banks and card networks to streamline the credit card dispute management process resulting in higher success rate of up to 60% from 35%. 

  • Implemented customized dispute management framework in collaboration with cross-functional teams engineering and product teams for new markets in China and India thereby saving $2.5M USD.

  • Led cross-functional efforts initiatives to identify, induct and standardize fraud data from third parties that resulted in a lift of 10-15% in identification of specific kinds of credit card fraud on the app store.

  • Implemented a Fraud early warning system using time series analysis to identify potentially active large scale Fraud rings across different markets.

  • Developed and operationalized a central fraud repository of labels from different sources which was then used for fraud model training and feature development

July 2018 to Aug 2020

Quantitative Business Analyst, Apple - Strategic Data Solutions

  • Led a project to expand existing process of that ensures more accurate actions by fraud prevention agents. The project saved over $1.6M for Apple. 

  • Partnered with Fraud prevention specialist to create a batch process with an increased throughput of 80% to identify fraud orders. 

  • Post success of the above partnership - onboarded, trained, and mentored 3 additional Fraud prevention specialists to create a new Technical Fraud Analyst team. 

  • Automated a batch action process that released least risky AOS orders on review which was equivalent to the work done by 3 full time fraud prevention specialists, with over 98.5% accuracy. 

  • Created a new reporting process using Tableau for multiple programs under AOS to provide insights into operational data daily and report on long term trends and metrics. 

July 2014 to May 2017

Assistant Manager, CitiBank, Bangalore                                                                                                                                                          

  • Led a team to Create and automate the delivery of a digital metrics repository of 400+ key process indicators for 20+ markets in partnership with data warehousing and digital banking teams. 

  • Identified 30,000+ unaccounted customers and 1.1 million Indonesian Rupiah missing deposits and balances when working with Citi Indonesia; oversaw implementation of corrective measures in their SAS based MIS reporting to rectify the process. 

  • Collaborated with Citibank Taiwan to negotiate between the local team and the regional team to resolve previously miscalculated and over-reported Credit Cards revenue by almost 500,000 USD. 

July 2014 to May 2017

Research Intern at Regional Integrated Multi-Hazard Early Warning Systems, Bangkok                                                    

July 2014 to May 2017

Research Intern at Indian Institute of Technology, Delhi, Dept. of Management Studies                                                                 

Education & Certifications

June 2024                           

May 2018

May 2018

Cornell University, Online, Machine Learning Certificate

Carnegie Mellon University, Pittsburgh USA, Master of Information Systems Management, Heinz College

Sardar Vallabhbhai National Institute of Technology, Surat, India,        Bachelor of Technology in Computer Engineering

Side Projects &
Technical Compentices

  • Short term stock market price classifier:

    • ​Implemented a short-term stock market price predictor by scraping data using yahoo finance, beautiful soup library. The project includes download and manipulation of data, training and analysis of model performance.

    • The implemented RF-Classifier model can predict with 91% precision on which stocks can go up in price by over 10%

​

  • Technical Skills:

    • Coding languages: Python (Pandas, Numpy & scikit-learn), SQL

    • Visualization tools: Python(matplolib), Tableau

bottom of page