Gaurav Garg Email & Phone Number

Gurugram, Haryana, India

Gurugram, Haryana, India

Gurugram, Haryana, India

- Product page Recommendations Recreated ETL pipeline increasing scalability and efficiency to daily recommend ~9lk products. Made improvements to pre-existing collaborative filtering based approach and added nudges to improve quality of recommendation. Achieved an incremental ~10% increase in recommendation coverage and ~1% improvement in conversion- Sales velocity Portal Created frontend and backend for a portal to disable/enable , assign restrictions on daily sales limit on select sellers for internal usage. Frontend - HTML, Bootstrap, Javascript Backend - Flask, MySql- Competitor price extraction Developed pipeline to extract and analyze product details of relevant competitors

Gurgaon, Haryana, India

- Price recommendation Used traditional machine learning techniques to recommend optimal prices for catolog with 80% coverage while maximising conversions and minimizing reduction in net margin. Recommended optimized ASP for different levels of product hierarchy . Developed framework for reinforcement learning based approach. Achieved significant increment in daily orders- Propensity modelling Developed model using traditional machine learning algorithms to predict 1. Organic probability of purchase by a user in near future 2. Probabilty of purchase after being marketed - Image deduplication Built a highly scalable ETL pipeline to daily analyze and assign groups/clusters to duplicate products with a catalog size of >10M products. Used scale invariant Feature transform techniques to great effect. Tried Image Hashing , transfer learning techniques as well . Achieved in flagging an incremental ~5% duplicate images - Image Quality Detection Built process to filter blurred and pixelated images using 1. Computer vision based NR - IQA techniques 2. Transfer learning based DNN techniques

Gurgaon, India

- Image Background Removal Used traditional Computer vision techniques , Image matting and transfer learning based approaches to segment images.

Bachelor Of Technology, Mathematics And Computer Science, 8.32

Computer Science, 95.4%