◆ Developed a tree based model to rank the ads showcased on the search result page. The model aimed to enhance the relevance and effectiveness of displayed ads , ultimately improving the overall user experience and engagement.◆ Developed a model to predict the impressions for Manual ads sellers.The real-time nature of the system allows sellers to make informed decisions on bid adjustments for optimal advertising performances.◆ Built an optimization Model to allocate the budget and… Show more ◆ Developed a tree based model to rank the ads showcased on the search result page. The model aimed to enhance the relevance and effectiveness of displayed ads , ultimately improving the overall user experience and engagement.◆ Developed a model to predict the impressions for Manual ads sellers.The real-time nature of the system allows sellers to make informed decisions on bid adjustments for optimal advertising performances.◆ Built an optimization Model to allocate the budget and bid to all products automatically for e- commerce marketplace sellers on a daily basis to maximise the orders and budget absorption. ◆ Deployed highly scalable and low latency model inference pipelines.◆ Integrated Vertex AI Feature Store functionality into the system, enabling efficient management, storage, and retrieval of features.◆ Developed Apache airflow pipeline to automate the end to end process of data generation , processing, model training and deployment , ensuring a streamlined workflow. Show less

▪Designed and implemented a specialised module utilising advanced Machine learning and Deep learning algorithms to analyse facial attributes such as face width and shape extracted from customer's photo. This innovative approach enables the system to generate personalised recommendations tailored to individual customers. ▪ The personalised recommendations have proven to be highly effective contributing to noteworthy 5% increase in conversion rates. Additionally… Show more ▪Designed and implemented a specialised module utilising advanced Machine learning and Deep learning algorithms to analyse facial attributes such as face width and shape extracted from customer's photo. This innovative approach enables the system to generate personalised recommendations tailored to individual customers. ▪ The personalised recommendations have proven to be highly effective contributing to noteworthy 5% increase in conversion rates. Additionally , the system has demonstrated a remarkable 12% reduction in return rates , showcasing its impact on enhancing customer satisfaction , acquisition and retention. Show less