A system that identifies and blocks the cash-on-delivery option for faulty customers with bad buying histories. Previously, we had no way of tracking the customer performance, which led to many customers rejecting the delivered orders at their doorsteps, causing Daraz to bear the failed logistics cost. This system enabled us to block a cash-on-delivery (COD) feature for certain customers and make them pay in advance for their orders. It is based on a delicate trade-off as it increases gross-to-net revenue but can also decrease the customer base due to the COD feature blocking for parcel deliveries. I first conducted a thorough data analysis to find the impact on the business and moved on to creating data pipelines and a performance dashboard that would gauge the impact of the system on the overall business of Daraz.

An automated alert system that notifies customers about delayed orders based on specific logistics metrics in order to enhance the customer experience. I worked on developing the system's end-to-end data pipelines, designed the business flow, and made a BI dashboard to gauge the performance.This project not only enhanced the customer experience but also helped in gauging Daraz's logistics performance and highlighted key metrics that needed to be fixed.

Every data visualization dashboard consumes a certain amount of computing and memory resources. Knowing how many resources the dashboards consume from the assigned cloud quota is imperative when working in the eCommerce industry. Currently, there are more than 700 dashboards in Daraz. When these dashboards are refreshed daily, they consume many resources, slowing down other processes. Therefore, I needed to identify which dashboards were the most frequently used and which were not so they could be decommissioned to save resources. I created a meta dashboard that would rank the dashboards by tracking the daily, weekly, and monthly active users and their visits. Also, this meta dashboard tracked individual user history on multiple dashboards, i.e., the number of dashboards that a particular user regularly visits, which helped us filter out the executives' dashboards.

At my previous company, Afiniti, multiple clients used the Afiniti engine to optimize their call center performance based on the data-driven decision-based customer and agent pairing. The legacy enterprise data portal that Afiniti used to gauge clients' performance had some limitations. For instance, there was no implementation of change data capture and historical analysis of the clients. Also, the optimizing metric, such as handle time, wait time, etc., that was used to calculate the performance of a client was not recorded historically. The enterprise data warehouse (EDW) structure caters to all limitations of an enterprise portal along with additional features, such as a standardized model that can fit into different business requirements without any change in architecture. It helped us track historical changes made to clients' performance and provided a holistic view of all clients in a single portal and at any time.I worked on creating the whole data warehouse from scratch, including developing all data pipelines and dimensional modeling.

A data integration pipeline that pulls data from certain data entities in Microsoft Dynamics 365 into our supply chain meta-model at Seeloz. I developed this data integration pipeline in Azure Logic Apps to fetch data from data entities and load them into Azure Blob Storage, which could later be used in ETL written at our end. All the communication was done using Azure Service Bus. The app was triggered using the HTTP POST request, and the required arguments were passed using the JSON payload. All the error handling and logging were also implemented adequately at each step.