Headline

Senior Data Scientist @ Walmart | Machine Learning | AI

About

Senior Data Scientist with more than 7 years of experience, currently spearheading analytics for an Ads Marketing program of a Big 5 Tech client. My areas of expertise include Classical ML, Deep Learning, A/B Testing, Credit Risk Analytics, Statistics and Mathematics. The idea of solving unique real-life problems with data excites me, especially if it requires me to go beyond black-box algorithms and create customized solutions tailored to the particular problem. As a technophile, I love working on personal projects in areas ranging from Machine Learning to Internet of Things (I have also published a research paper on robotics/IoT in IESRJ journal). Check out some of my projects on my GitHub profile: https://github.com/nandykaustav. Most of these are passion projects and not simply done for the sake of building a portfolio, so I would love any feedback or discussion about them.

Location

Bengaluru

AI stack

anthropic apis

cursor

hugging face

langchain

langgraph

openai

Work experience

Senior Data Scientist at Walmart Global Tech India, Jan 2025 - Present · 1 yr 8 mos, Bengaluru, Karnataka, India, - Rebuilt item recommendations for Walmart sellers using a two-tower architecture (TFRS) recommender system, suggesting items from a list of 1.4 million items to 3P sellers which resulted in a GMV increase of $9 million USD. - Built a multi-agent AI system using LangGraph for the seller onboarding team, that orchestrates the first-sale enablement workflow. Autonomous agents read seller data, identify listings close to conversion and recommend interventions to help sellers achieve their first sale using a seller activation model and DiCE, and generate contextualized outreach emails for each seller. - Implemented the price optimisation module for Walmart’s 3P sellers that includes price elasticity modelling using Bayesian hierarchical model and optimization using NSGA-II optimizer to suggest optimal discounts on items to maximize GMV gain, which achieved a GMV gain of $14 million USD for Mexico alone. - Revamped the offer quality model within 1.5 months of joining, enabling the removal of 3M low-quality offers from the platform.

Senior Data Scientist at Fractal, May 2022 - Jan 2025 · 2 yrs 9 mos, Spearheaded the data science driven growth of a Big-Tech client's Ads Marketing & CSR programs as a Marketing Data Scientist. Won the Star Award in Client Champion category for helping in growing overall reach by 42% within my first year on the team. - Drove up email open rate and click rate by 23% and 131% respectively by building and deploying in GCP a user behaviour classification model for email audience segmentation using a stacking ensemble of Logistic Regression, SVM, Neural Network & XGBoost. Implemented model monitoring pipelines to monitor the model continuosuly. - Reduced 150 hours of work per quarter by developing a custom fuzzy algorithm for automating the previously manual process of matching user-input company names with internal company list, based on a modified version of Levenshtein distance. - Increased 6-month average reach by 19% using a Random Forest Regressor to predict lifetime reach of facilitator candidates. - Forecasted program KPIs for the following year using an ensemble of SARIMA and Prophet time series models. - Achieved a 22% lift in conversion rate for new and 15% lift for all facilitators by recommending operational strategies based on actionable insights derived from the results of A/B experiments designed to pinpoint optimal values of factors affecting KPIs. - Built/Maintained 14 organization-wide dashboards using Data Studio/Looker to present the data about the program to stakeholders across the company. - Created ETL/MLOps pipelines to deploy and monitor models on Vertex AI using GCP services like BigQuery, Cloud Functions etc. - Saved the project at least £50k in outsourcing costs by building cost-effective in-house automation tools for processes like sending bulk emails, generating customized certificates for facilitators etc. using Apps Script (Javascript).

IT Analyst ( Data Scientist ) at Tata Consultancy Services, Mar 2017 - May 2022 · 5 yrs 3 mos, Kolkata, West Bengal, India, - Developed and optimized 2 credit risk models to predict Probability of Default (PD models) from scratch for retail lending segments of a fintech startup for financial risk assessment purposes. - Created a dynamic price optimization model for an e-commerce company. Used K-Means Clustering to group the customers, ran a price elasticity model for each group and optimised revenue to give targetted discounts using constrained optimisation. - Led a team of 3 junior developers to implement an automated end-to-end Machine Learning platform for data mining, exploratory data analysis (EDA), feature engineering, model training(Logistic Regression, SVM, Decision Trees, Random Forest, Neural Networks, K-Means Clustering) and model evaluation using Scikit-Learn and TensorFlow. - Implemented both frontend and backend structures for a module in a financial product suite by TCS. Fetched data from bank’s middleware and database and wrote scalable backend APIs using Flask, while working under Agile methodology.

Assistant System Engineer at Tata Consultancy Services, Mar 2017 - Jan 2022 · 4 yrs 11 mos

System Engineer at Tata Consultancy Services Limited, Mar 2017 - Jan 2022 · 4 yrs 11 mos, Greater Kolkata Area

Education

Bachelor of Technology (BTech), Electronics and Communications Engineering, institute of engineering and management

High School, Science, Ram Mohan Mission High School

Public links

github: https://github.com/nandykaustav