Q: With the emphasis on decentralization, how do you manage data platforms?
A: Despite our inclination towards decentralization, data platforms should remain central. Governance is crucial. Having dedicated personnel oversee these platforms ensures strategic and long-term vision, preventing the adoption of fleeting technological trends.
Q: When establishing strategic partnerships with tech companies, what criteria do you consider?
A: Two primary factors influence my decision. Firstly, it's the people. Trust is paramount. I assess the potential for a productive business relationship and gauge the longevity of the partnership. Secondly, it's the technology. It's essential to trust the technology they employ and their approach to problem-solving. For instance, while I'm a proponent of decentralization, platforms supporting decentralized teams should be centralized.
Q: What role does user-centric design play in platform success?
A: It's imperative to always focus on the user, aiming to simplify their interaction with the platform. The product manager plays a pivotal role in this, especially those with significant experience in data platforms. Integrating "glue code" can serve as a bridge between various tools, offering users a unified interface and reducing the necessity to deeply understand each tool individually. Without strong product management, platforms become collections of tools rather than cohesive products.
Q: Are engineering best practices architecture-dependent?
A: No. Good engineering practices can be implemented irrespective of the chosen architecture. These practices, which range from being data-driven to efficient incident management, are foundational and don't rely on specific technologies. The platform engineering team, given the critical nature of their work, often pioneers the adoption of best practices and serves as evangelists for these standards across the organization. DevOps, infrastructure as code, and standardized deployment pipelines are essential regardless of whether you're running a data warehouse, data lake, or data mesh.