Mar 2022 – Jan 2023 · 10 mos · Chicago, Illinois, United States

• At ZS, I led various data science projects, working alongside teams to implement advanced methodologies:
• Primarily worked on researching Trustworthy AI frameworks and global AI regulations, using OpenAI's GPT3 API to distill extensive regulatory documents, contributing research to notable conferences like NeuIPS.
• I also Conducted research focused on detecting biases in machine learning models to assist clients in ensuring fairness in their AI systems.
• Oversaw projects involving the creation of predictive models and simulations for a CPG brand, aiding in the design of new product concepts.
• Guided a team to develop customer-lapse models using Deep Learning LSTM for an insurance company
• Tech Stack: AWS, Power BI, Azure ML, Azure Synapse, Microsoft suite
• Data Science: GPT3, Random Forest, XGBoost, CatBoost, Long Short Term Memory Networks, IBM’s AI Fairness 360, Google’s What-If tool
• Language/Scripting: Azure Synapse SQL, Python

May 2021 – Jan 2022 · 8 mos · Bengaluru, Karnataka, India · Remote

• At Twitter, I served within the Trust team, focusing on ensuring Account Content Integrity and operational efficiency. My responsibilities entailed collaborating with operations teams, strategizing to enhance agent efficiencies, and supporting financial and strategy teams through time series modeling for effective planning.
• Tech Stack: Presto, Postgres, AWS, Google suite
• Data Science: Random Forest, Logistic Regression, Isolation Forest, SARIMA
• Language/Scripting: Presto SQL, Python(Pandas, numpy, matplotlib, sklearn, pyod, statmodels)

Mar 2020 – Apr 2021 · 1 yr 1 mo · India

• Helped Bajaj-finserv App campaigning team with insights to improve Click-Through-Rate (CTR) metric using Random Forest and decision tree techniques
• Re-defined the Digital Marketing process by personalizing the creatives to each customer micro-segment using pillow-python library, opencv, yolo pretrained object detection model and Random Forest techniques. The A/B tests showed that, the new process led to an increase of Click-Through-Rate (CTR) from 1.2% to 1.8% for high value customer segments of the company
• Created power-BI governance dashboards to optimize push-notification strategy for Bajaj Mobile-app
• Tech Stack: Azure Synapse, Databricks, Data Factory, PowerBI, Microsoft suite
• Data Science: Random Forest, Object detection(opencv), Dashboards, Insights
• Language/Scripting: SQL, Python(Pandas, numpy, matplotlib, sklearn, opencv, statmodels)

May 2019 – Feb 2020 · 9 mos

• Helped the content management team of Target.com to understand the main drivers of traffic conversion in high ticket accessories section using Random Forest and linear regression (after binning the variables to handle non-linear relation with conversion rate)
• Created an automated k-medoid clustering web-application to cluster given Target stores based on user input clustering variables
• Helped the HR-operations team to understand the drivers of Team-Member attrition in Target stores using Linear regression model, which resulted in tuning of the most important factors to reduce Team-Member attrition from an average of 65% to 58% YOY
• Tech Stack: Hive, Teradata, Tableau, Domo, Linux machines, Microsoft suite
• Data Science: Random Forest, K-medoid, Linear Regression
• Language/Scripting: SQL, R, Python(Pandas, numpy, matplotlib, sklearn, Django)

Nov 2015 – Apr 2019 · 3 yrs 5 mos

• Created excel-dashboard for store-staff hiring plan which was fed by ARIMA forecasts of staff vacancies and hiring costs from various hiring pools in the market. Automated R-code for forecasting eliminated ~14 man-days per quarter manual efforts and hiring experts could focus more on trying different simulations to optimize hiring costs

Nov 2014 – Nov 2015 · 1 yr · Bengaluru Area, India

• Created Failure prediction model for in-car advance warning systems for a leading British multinational automotive company
• Led a team of 4 members to translate the then existing MATLAB Simulink self-learning car algorithm into python and finally implement the algorithm for in-car demo to Jaguar dealers across the globe
• Tech Stack: Hive, Tableau, Linux machines, Microsoft suite
• Data Science: Random Forest, Logistic Regression, Association rule mining
• Language/Scripting: SQL, Python(Pandas, numpy, matplotlib, sklearn, Django)

Mar 2012 – Oct 2014 · 2 yrs 7 mos

• Helped in building auto ML tool to Measure the impact on sales from events like promotions, merch resets etc., for a large US retailer
• Identified drivers of customer churn using Logistic regression for a leading electronics retailer in the US
• Helped in Next transaction date prediction of customer, using recency as a dependent variable against previous transaction gaps and other demographic & seasonality variables as predictors