Mohammed Kajee

01 Building the Infrastructure

Every data career starts at the bottom of the stack: the databases, the ETL pipelines, the warehouse models. Mine started in banking, where data volumes were measured in terabytes, tolerance for errors was zero, and if a pipeline failed overnight, hundreds of branches didn't get their reports in the morning.

I built data warehouses using both Kimball and Inmon approaches: star schemas, dimensional models, the foundational patterns that turn messy operational data into clean analytical structures. The platforms spanned Teradata, Microsoft PDW, Greenplum, and PostgreSQL. As data volumes grew beyond what traditional warehouses could handle, the work expanded into the big data ecosystem: Hadoop and Hive for batch processing, Kafka for streaming. The platforms kept changing, but the job stayed the same: build the infrastructure that turns raw data into something an organisation can actually use.

In one engagement, I inherited performance issues on 2TB+ databases serving a 600+ branch banking network and built monitoring solutions that recovered approximately 6 hours of daily processing time. That was when I understood what data at scale actually meant.

02 Turning Data into Decisions

The next phase was about what happens after the infrastructure is built: turning massive data into intelligence. I worked across Power BI, QlikView, and Tableau, building the reporting and dashboarding layers that put data in front of the people who needed it. In retail, that meant making 1,800+ stores worth of transactional, loyalty, supply chain, and inventory data legible to decision-makers. In the tech sector, it meant designing BI architecture for a global portfolio spanning media, e-commerce, fintech, and classifieds across dozens of countries.

Most companies have plenty of data. Very few have genuine intelligence. The hardest part was never the technology. It was bridging the gap between what an organisation collects and what it actually uses to make decisions. That became the work.

03 Data as a Product

E-commerce taught me to think about data differently. When supply chain operations, customer analytics, merchandising, and fulfillment all depend on the same data platforms, you stop thinking in terms of reports and start thinking in terms of products. Cloud-native platforms like BigQuery became the default. The shift wasn't just technical. It was a mindset change from serving analysts to serving the entire business.

Building data products means owning reliability, not just accuracy. If the pipeline breaks, it's not just a dashboard that goes dark. It's operations, logistics, and customer experience that feel the impact. That kind of accountability changes how you build things.

04 Strategy, Architecture & AI

Now I lead analytics and strategic data projects at one of South Africa's Big 5 banks. After 15 years across banking, retail, tech, and e-commerce, I've seen what good data architecture looks like in very different environments, and I know which patterns transfer across industries and which don't.

The newest chapter is AI. I'm enrolled at Udacity's Institute of AI & Technology, building AI agents, and shipping open-source projects. After 15 years of building the data infrastructure that AI depends on, the transition isn't a leap. It's the next layer of the same stack.