On-Demand Webinar · 1 hr 1 min

The Role of DataOps in Data Modernization

Cognizant's Jayaprakash 'JP' Thakur, Senior Director and Head of Data Modernization, joins DataKitchen's Chris Bergh on why DataOps belongs at the center of a modernization effort rather than bolted on at the end of one: continuous delivery of data and insight, faster transitions to new data ecosystems, and the foundation any agile data architecture needs. Recorded June 2021; updated August 2026.

Presented by Chris Bergh

What you'll learn 7 points
  • Cognizant reports that 90 percent of its data modernization customers are asking for CI/CD and DataOps pipelines, while roughly 56 percent of client data teams still manage their data pipelines manually.
  • Cognizant's DataOps work with a leading banking service provider cut cycle time from 12 weeks to two weeks by automating release orchestration, deployment, integration, and continuous testing on Pivotal Cloud Foundry.
  • A Cognizant financial services engagement onboarded 600 application projects onto an integrated CI/CD stack with GitHub and Jenkins, ran more than 100 application builds and deployments a day with no manual intervention, and increased release frequency by 35 percent.
  • Cognizant frames the problem of today's data and analytics ecosystem as the three Cs: complex, confusing, and costly, with too many tools, frequent pipeline changes, and people spread across the organization.
  • DevOps and workflow tools fall short on data work for six reasons: no end-to-end meta-orchestrated production pipeline, no environment pipeline, a process that is not DevOps CI/CD, the team and data center coordination that data analytics requires, no common system and vocabulary, and no process measurement to drive behavior change.
  • Moving to AWS, Azure, or GCP delivers a powerful collection of tools with no defined process for using them as a system. That is data integration without process integration, and it leaves the customer to design the DataOps superstructure themselves.
  • At a top 10 global health company, schema drift between an on-premises analytics team and an Azure cloud analytics team caused delays, errors, and team conflict until schema management, meta-orchestration, testing, monitoring, and versioning were put in place.

Slides

47 slides

Questions from this session

What is DataOps?

DataOps is the practice of automating the orchestration of a data pipeline through acquisition, integration, transformation, and consumption, so data is delivered and monitored continuously with agility and assurance. Gartner describes the goal as delivering value faster by creating predictable delivery and change management of data, data models, and related artifacts. Just as DevOps brought development, QA, and operations together, DataOps brings the data stakeholders together.

What role does DataOps play in data modernization?

Modernization projects replace the platform but usually carry the old process with them: errors in dashboards and pipelines, slow deployment of new features, poor collaboration across distributed teams, and no measurement of productivity or SLAs. DataOps addresses those four upstream processes directly, so a move to the cloud reduces error rates and cycle time rather than reproducing them on new infrastructure.

Why do DevOps and workflow tools fall short on data projects?

Six gaps. There is no end-to-end meta-orchestrated production pipeline across the toolchain, and no environment pipeline. The process is not DevOps CI/CD. Data analytics requires team and data center coordination that workflow tools do not model. There is no common system or vocabulary across teams. And there is no process measurement to drive a change in behavior.

What results have DataOps modernization projects produced?

Cognizant reports a leading banking service provider cutting cycle time from 12 weeks to two weeks, a financial services provider running more than 100 automated builds and deployments a day across 600 onboarded application projects with a 35 percent increase in release frequency, and a life sciences provider projecting a 40 percent effort reduction on Azure.

What tests and monitors reduce production data errors?

Five test types carry most of the load: traditional data quality checks, statistical process control, location balance tests, historic balance tests, and business-based tests. They have to run automatically in production, sit on top of the entire toolchain rather than one tool, send alerts, keep history, and be easy to create.

What is meta-orchestration?

Meta-orchestration runs an end-to-end pipeline across the tools that each already have their own scheduler, spanning the database, ETL, BI, data science, and governance layers and spanning development, test, and production environments. It sits alongside storage and version control, history and metadata, permissions, environment secrets, and automated deployment as part of a data architecture built for change.

Where to go next