Ali Ashfaq

Built the core modules of one of the biggest eCommerce systems in the USA.

The project used analytics and business intelligence (BI) to improve customer experience and engagement, retain potential customers, offer them the best choice, enhance the system's efficiency, and increase sales based on historical and incremental data.

I designed and built multiple end-to-end ETL processes for data ingestion and transformation in the Google Cloud Platform (GCP) and coordinated tasks among the team. The core modules allowed the system to integrate data from various heterogeneous source platforms, including Google Cloud, Amazon S3 bucket, MuleSoft API, and SFTP. It also featured the ETL process by incorporating data in the GCP deep learning and BigQuery DWH and ingesting the data in the Neo4J graph database.

I used Python, SQL, and GCP stack for data processing and Apache Airflow for data orchestration during the project. I used the Google Cloud function with Python to load data into BigQuery for on-arrival CSV files in the Google Cloud Storage (GCS) bucket and maintained raw file archival in the cloud storage.

Created a credit scoring model to streamline the bank's loan approval and dispersing processes, which increased revenue by $500 million.

My main contributions included the ETL process, EDW architecture, and aggregated data set (ADS), using technologies such as Vertica, IBM InfoSphere DataStage, and IBM Cognos.

I developed the model's architecture that was successfully launched with over 90% accuracy and efficiency. It features the ETL process of extracting loan/lease transactional data from 70+ financial institutions and loading the transformed data into the DWH.

I also prepared an ADS to aid the formation of the statistical model. It can predict and categorize the customer/borrower as either good, average, or bad and assign a default probability to every customer based on the information provided to the model.

I empowered Locallogy, a small business digital marketing agency, with a comprehensive data warehouse solution that revolutionized its in-depth analysis capabilities. This solution enabled me to consolidate and analyze data from multiple sources, providing powerful insights through a unified dashboard.

To achieve this, I designed a robust data flow incorporating data from various APIs, such as Screaming Frog and AWR (Advanced Web Ranking). I created a potent combination of data sources by integrating Google Analytics and Google Search Console with BigQuery. Additionally, I implemented an interface on Google Sheets, seamlessly connected to BigQuery, facilitating collaboration across the organization. This adaptable solution allowed the team to efficiently insert, update, and delete data from BigQuery, enhancing their ability to deliver robust digital marketing solutions for their clients.

From a technical perspective, I leveraged several Google Cloud Platform (GCP) services, including Compute Engine, Cloud Scheduler, and BigQuery. I employed VM instances running Python ETL scripts to capture and process data, ensuring seamless integration with the APIs.

I significantly enhanced the performance of an intermediary company by delivering a robust business intelligence solution.

By implementing an end-to-end data pipeline, I successfully gathered, transformed, and loaded data from various sources, such as Heroku, Stripe, HubSpot, Toggl, and Google Sheets, into a centralized storage system BigQuery Data warehouse. The system included comprehensive dashboards tailored for executives, project managers, team leads, and recruitment departments, enabling them to make quick, informed, data-driven decisions.

I used DBT and SQL in this project, integrating with GitHub. To accommodate the substantial volume of daily data generated by over 5000+ projects and 1000+ developers, I designed and implemented efficient ETL (Extract, Transform, Load) jobs.

These jobs seamlessly combined data from multiple sources, resulting in a unified and reliable source of truth for the organization. Through this implementation, the company gained a comprehensive and streamlined data infrastructure, providing accurate insights for enhanced decision-making.

Technically, the core data model was designed using DBT, GCP Bigquery SQL, and Looker.

The M&M Data Warehouse architecture was developed to optimize the integration of large datasets extracted from multiple heterogeneous source systems. This architecture aims to enhance the efficiency of the decision-making process for Marcus & Millichap's legal property system in the United States.
By streamlining the information flow, it enables better insights and more informed decision-making. The M&M Data Warehouse architecture encompasses multiple areas, including the extract, transform, load (ETL) process, which ensures seamless data incorporation into the warehouse after transformation and cleaning.

As a senior data engineer for Eu.drsturm.com, an online cosmetic retailer operating in Europe and America, I spearheaded the integration and analysis of complex data sets from Clickstream into BigQuery. My role encompassed architecting and executing a robust ETL pipeline to efficiently transport data, utilizing Dataproc for advanced processing and optimization tasks. I was instrumental in transforming raw data into actionable insights, which were visualized through Looker Studio. This enabled the visualization of critical business metrics, including sales performance, voucher redemption rates, and the effectiveness of promotional campaigns. My contributions were pivotal in enhancing data-driven decision-making, optimizing marketing strategies, and ultimately driving the company's growth. My work showcased not only my technical expertise but also my ability to collaborate closely with cross-functional teams to align data analytics with business objectives.