Apr 2025 – Sep 2025 · 5 mos

• I am currently leading the Demand Decomposition initiative for Sam’s Club, developing a unified ML/Deep Learning–based forecasting solution that breaks down item-level demand into baseline and driver components such as promotions, weather, and out-of-stock events. The system integrates diverse datasets including sales, inventory, promotional signage, weather, and OOS data to deliver accurate forecasts and actionable insights.
• This platform would enable business teams to quantify the impact of external drivers and work on improvements like optimizing promotional strategies, improve availability planning, and enhance forecasting accuracy across both O and S status items.

Aug 2023 – May 2025 · 1 yr 9 mos

• Tredence is a service-based company. As a Data Scientist here, I work on delivering AI solutions to Retail Clients. As part of my time here I have spent a lot of time tailoring solutions for Sam's Club. I have,
• 🚀 Redesigned and scaled Sam’s Club’s AI-based E-commerce Review Summarization pipeline, transitioning it from real-time to a robust batch processing framework, boosting system efficiency, stability, and scalability for large-scale customer feedback analysis.
• 🧠 Developed a machine learning–powered Buy Quantity Optimization engine for 10,000+ seasonal and One Time Buy (OTB) SKUs across 400+ locations leveraging demand forecasting, markdown strategy, and price elasticity modeling to unlock six-figure cost savings per club.
• 📊 Built and deployed a full-stack AI-driven Stock Level Prediction WebApp, enabling near real-time visibility into in-season sales trends, reducing revenue loss due to out-of-stock and overstock scenarios across merchandising teams.
• 🤖 Designed a GenAI-based Natural Language Query (NLQ) to BigQuery interface for Walmart Canada—translating user queries into SQL/dataframes with >90% relevance, powered by LLMs and deployed on Google App Engine to accelerate data access for business users.

May 2022 – Jul 2022 · 2 mos · Chandigarh, India

• Developed a trajectory mapping system for blast shockwaves using image processing and DL, enhancing research efficiency.
• Designed and fabricated a 90% cost-effective high-speed counter to measure impact velocity of striker bar in Split Hopkinson’s Pressure Bar setup.
• Tools Used – SolidWorks, Arduino Uno, TinkerCAD, C++

Dec 2021 – Jan 2022 · 1 mo

• Gained knowledge of the ONGC’s process of Oil Well Drilling and got familiarized with the work culture in a PSU.

May 2021 – Apr 2022 · 11 mos · India

• ‘Designed a humanoid actuated by Mckibben’s artificial muscles’. In this project, I contributed to the design of the model in SolidWorks 2019 and performed the numerical analysis of the working of muscles in Ansys and MATLAB. Further, I developed the mathematical model to execute the gait cycle for walking straight by deriving particular angles using pressure change in muscles and coded the same in MATLAB. At last we did the sensor-based simulation and PID tuning of the whole structure using the Robot Operating System (ROS).