Shubhanshi Sharma

AI Researcher

Gurugram, Haryana, India5 yrs 3 mos experience
Most Likely To SwitchAI ML Practitioner

Key Highlights

  • Increased model accuracy by up to 50% at American Express.
  • Developed forecasting models for 3K distribution centers.
  • Delivered a chatbot resolving 50% of customer service queries.
Stackforce AI infers this person is a Data Science professional with expertise in Fintech and Logistics.

Contact

Skills

Core Skills

Statistical ModellingData AnalysisMachine LearningData ScienceForecastingPerformance TrackingAi DevelopmentCustomer Service Automation

Other Skills

AIMLARIMABusiness AnalysisC (Programming Language)Data GovernanceData VisualizationDemand ForecastingGoogle AnalyticsLGBMLight GBMMySQL WorkbenchPower BIPresentation SkillsProblem SolvingPython

About

👉🏻 Committed and result-oriented individual with solid 4+ years of experience in the Financial & Logistic organisation and academic preparation holding a Master’s degree in Applied Statistics and a Bachelor’s in Mathematics. 👉🏻 Proven ability to assess and manage complex unstructured datasets, providing valuable and actionable insights. 👉🏻 Viewed as a strong communicator, decisive team leadership and visual storyteller. 👉🏻 Proficient with Statistical Modelling, CRISP-DM Methodology, Data Governance, gathering and pre-processing skills. 👉🏻 Technically profound with Python, MySQL Workbench, Google Analytics, R, Power BI, Tableau.

Experience

American express

Senior Analyst - Data Science | Credit and Fraud Risk Finance

Apr 2022Present · 3 yrs 11 mos · Gurugram, Haryana, India · Hybrid

  • Member of the CECL Reserve Analytics team - Responsible for modelling and analyzing trends in AXP’s credit reserves.
  • Took over ownership of the reserve models for Personal and Small Business Loan portfolios. Identified flaws in the existing model and formulated enhancements and solutions that increased model accuracy by up to 50%.
  • Developed model monitoring methodology to accurately evaluate reserve forecast against actuals using the MAPE metric, which was up to 40% more accurate than the previous methodology.
  • Led the creation of a new reserve model and methodology for Small Business and Personal Loans, a $10B+ portfolio.
  • Extensive experience working closely across departments, including governance, product, risk, pricing, and strategy teams.
  • Foundational knowledge and exposure to PD, LGD, and EAD models through involvement in CECL-related initiatives.
  • Collaborated on the comprehensive CECL quarterly reserves booking process, leading efforts in data optimization, visualization, and presenting QoQ decomposition insights to senior stakeholders
  • Developed and automated a decision tree classifier model to predict customer propensity for Business Loans, optimizing performance using GINI for effective targeting.
  • Led key strategic initiatives including Portfolio Allocation, FPO Trigger Campaign, and SBS Pre-Attrition Campaign, driving customer targeting, spend growth, and attrition mitigation through data-driven insights and execution.
Statistical ModellingData GovernancePythonMySQL WorkbenchGoogle AnalyticsR+3

Ecom express private limited

Associate Data Scientist

Aug 2020Nov 2021 · 1 yr 3 mos · Gurugram, Haryana, India

  • Demand Forecasting: Build “Forecasting models” using, ARIMA, SARIMAX, and Light GBM boost focusing mainly on aligning festivals, sales, holiday peaks, other influential factors, end -to end validation using Z-test and scaled up for 3K distribution centers with a projected profit of 4-5 Cr/yr, to minimize cost and, reduce breaches by allocating the optimum number of workforces per the forecast obtained.
  • Delivery Conversion Performance: Developed the quantitative indicators that helped track FE progress based on delivery conversion rate, fuel efficiency, attendance, and CSAT score, using an LGBM model and weighted scores.
  • Chatbot: Delivered a Menu-driven/ Rule-based Chatbot that resolved 50% of all customer service operations, using Python and AIML (Artificial Intelligence Markup Language).
  • Product Classification: Worked on ‘’Product Categorizations”, to map the product description provided by the vendor & the organization taxonomy (categories list) to reduce the risk mitigation.
Demand ForecastingARIMASARIMAXLight GBMPythonAIML+2

Comparepolicy.com

Data Analyst Intern

Apr 2019May 2019 · 1 mo · Noida Area, India

  • Analyzed the website traffic activities and suggested appropriate solutions to change marketing strategies, targeted platforms, audience resulting in a higher conversion rate than prior.
  • Performed Lead Conversion Gap – Funnel analysis and Identifying the potential customers (predictive analytics)

Education

Symbiosis Statistical Institute, Pune

Master of Science - MS — Applied statistics

Jan 2018Jan 2020

Banasthali Vidyapith

Bsc — Mathematics and Statistics

Jan 2015Jan 2018

Stackforce found 100+ more professionals with Statistical Modelling & Data Analysis

Explore similar profiles based on matching skills and experience