Ishaan Bassi

Whatsapp Recommendation: Generated 40% more EOIs than the existing algorithm in new homes business segment; Trained a random forest model based on 25 key features from user clickstream; Scaled the model to multiple whatsapp campaigns including reminder and instant messages; A unified model was created for resale and new homes business segments using 30 key clickstream features, boosting EOIs by 10% and cutting messages by 10%. Onsite Property Recommendation: Increased overall EOIs by 25% in the resale business segment; Developed a hybrid model by combining two recommendation approaches (content and collaborative); Enhanced the collaborative model using co-occurrence based similarity between various attributes of a property; Scaled the model to multiple touchpoints including search page, property details page and home page. App Based Push Notifications: Brought an overall jump of 50% in EOIs obtained from 99acres app notifications; Recommendations were generated using collaborative approach and packaged as per user intent and activity; Tuned the timing and number of notifications for best possible experience; Scaled the model to include entire audience after rigorous A/B testing against the existing algorithm.

Property Registry Data Pipeline: Created a full pipeline for crawling, parsing, mapping, and delivering Real Estate Registry Documents, offering price insights and trends to visitors on 99acres’ project and locality pages.

Topic Modelling of Biomedical Research Papers: Developed a model for categorizing and summarizing various biomedical research papers using Latent Dirichlet Allocation (LDA) algorithm in combination with Named Entity Recognition for enhancing the quality of dataset.