• Implemented STS (sales to subscriber) Forecasting (95-98% accuracy) & Target setting engine (Channel wise STS, Gross Add & hygiene KPIs) for more than 400+ distributor which helps to plan & drive DSTR accordingly.• Lead network monetization project to build an evaluation framework of 1440Mn BDT yearly investment on network expansion and to gain 900Mn+ yearly incremental revenue.• Leveraged R & Excel to develop a UI for Amar Khushi (Retailer self-campaign on Retail App) campaign configuration resulting in 0% manual dependency.• Developed AI//ML based gamified campaign for retailers in Retail App.• Automated excel reports in R which saved 20 minutes of 40+ resources every month.• Reduced ERS lifting process time from a day to below 1 minute & remove physical cash handling using ML based Distributor daily ERS Purchase Order suggestion.• Removed 4 manual reporting by developing end-to-end automation process using Telegram Bot & R which saved 8 hours daily.

• Designed & implemented GP’s first ever Daily ERS Suggestion System for 450K+ retailers that reduced daily Out of Stock from 6% to less than 2%.• ERS Suggestion system helped to maintain POS level 3 days safety stock coverage among follower POS and Distributor.• Developed an ML based personalized backend of first ever gamified self-managed trade program for retailers named “Amar Khushi” on 7+ kpis in retail app. Monthly more than 60% POS are utilizing this campaign.• Established retailer wise personalized product target engine for drive day which saves 20-30% commission in each drive than traditional product drive ensuring incremental revenue.• Measured brand visibility using ML & OCR of POS image.• Web scrapped Bioscope (Digital App) content to build a ML based content recommendation system to increase user retention.• Guided SE & POS’s daily journey through personalized instruction in Cockpit SE (Sales Executive mobile App) which completely removed paper-based information dependency.• Retail credit rating model.• Development of NRT Retail commission database system.• Development of distribution exception system.

• Achieved 85% accuracy using ML model to identify the gender of the subscriber using the scan copy (Image) of the subscription form where subscribers filled a gender declaration box.