Balaji Paraman

Built an end-to-end demand forecast pipeline(data-to-delivery) that provided weekly forecasts that helped the client to optimize their inventory allocation by reducing stockouts by more than 30%. Developed an end-to-end demand forecast pipeline(data-to-delivery) that provided long-range forecasts which helped the client better optimize their in-season and pre-season planning. Worked on a markdown optimization tool which helped the client to generate optimal recommendations on markdowns and clear excess inventory that helped generate around $3M USD in annualized margin impact. Developed an end-to-end demand forecast pipeline(data-to-delivery) that provided forecasts for the assortment optimization tool, resulting in a significant increase in accuracy for assortment planning considering planned sales.

Developed a model to accurately forecast demand for new products which improved accuracy by 25-30% compared to traditional ML models. Worked on developing a clustering pipeline for store clustering that finds the best algorithm-cluster combination for any given dataset. Developed and implemented a clearance markdown solution that leveraged demand transfer, base forecasting, and price-elasticity models that yielded a lift of 2% in revenue. Developed an automated business insight solution, that provided actionable insights for price and inventory recommendations. Improved the automated business insight solution by identifying and fixing critical bugs, increasing test coverage by 10%, and smoothing out QA functional testing through the addition of unit test cases and a logging module. Efficiently migrated a key use case from Python to SQL, resulting in a significant 89% performance boost for the business insight solution. Scraped holiday events data by country from the web to be used as part of forecasting project to predict the demand for personal computers.