Shrikrishna Padalkar

Senior Data Scientist

Capillary Technologies•  June 2022 - August 2024•  Remote

• Created notebook automations for various reports that are scheduled on a weekly and monthly basis using Databricks. This caters to the various Campaign/Leads requirements of the clients. • Analyzing the Leads for HP across states and zones. • Transaction Propensity Model:- Predict the probability of a customer transacting within next 30 days of his purchase. The solution was able to capture 70% (Gain) of the potential customers by targeting 50% base. A logistic regression model was trained after data pre-processing and feature engineering. The predictions on online data are being tracked via a ML pipeline. Tools:- SQL, Databricks, Python • Win-back Model:- Predicting the probability of a user transacting within next one year after his purchase. The solution was able to capture 69.5% (Gain) of the potential customers by targeting 50% base. Techniques used:- WOE and IV, Logistic Regression. Tools:- SQL, Databricks, Python

Data Scientist

ICICI Lombard•  June 2019 - October 2021•  Mumbai

Created a ML solution to identify potential claim rejections. The model was able to capture 84% of actually rejected claims. The model was trained on historical data of around 96K customers. Deployment:- Azure ML Studio. • Worked on identifying personal accident fraud claims based on historical data. 91% of the identified fraud cases were actually fraud. The model was trained on a small data set comprising approx. 2K records. • Curating the deployed models. Created automated mail triggers for capturing the day-to-day prediction results obtained via the deployed model. Tools:- Python, Azue ML Studio

Data Analyst

Findability Sciences•  April 2018 - July 2019

Day to day work comprises predicting market prices of commodities (Aluminium).