(US and Canada) Former Senior Director of Data Engineering, BI and Analytics for Petco, Kiran Kanetkar, speaks with AtScale CEO David Mariani about organizing teams, prioritizing data analytics goals of the organization, delivering data products, and modernizing data and analytics infrastructure in the cloud.

Speaking about organizing teams at Petco, Kanetkar says that his previous organization follows an agile model and has a cross-functional squad that includes data engineering resources and even analysts from the business side. He adds that, at times, the team gets additional support with either staff augmentation or managed services from external consulting companies since the skills are scarce, and the organization has to use resources wisely and effectively.

Recollecting how the company prioritizes its data goals, Kanetkar says that Petco has a goal set at the company level and is managed closely. He continues that the teams also have biweekly prioritization meetings with the business teams to ensure that the right priorities are identified and addressed. Kanetkar reveals that the team follows an agile sprint-based approach where they work for two weeks based on priorities set.

Sharing his view on delivering data products, Kanetkar says that treating data as a product is the right way to go, and that data-as-a-product can mean different things to different people. The teams can work with different business stakeholders as long as they have a clear definition of the data product they are supposed to deliver.

Kanetkar goes on to speak about important trends in data and analytics. He mentions that as more and more data is being analyzed, there will be newer applications powered by data and analytics initiatives. As traditional applications cannot be changed overnight, the organization might have to start building custom applications that rely on modern AI-ML algorithms, he suggests.

Sharing tips to modernize data and analytics infrastructure in the cloud, Kanetkar encourages organizations to re-engineer some of the ways they process data into the cloud-based data warehousing platform. He stresses that this is necessary because the old ways that might have worked on an on-premise enrollment will not work with the cloud, requiring organizations to put a lot more automation in place.