← Back to all talks

PlatformCon 2022 lightning talk: How we improve the Developer Experience at ClarityAI

EN
Year: 2022 Event: PlatformCon Core Talk

Description

This talk demonstrates how ClarityAI scales high-performance teams by treating its internal platform as a product, leveraging Team Topologies to reduce developer cognitive load and transforming manual release processes into automated, "fire and forget" workflows.

🎯 Key Learning

Scaling a high-performance engineering organization requires treating the internal platform as a product that prioritizes the reduction of developer cognitive load and the optimization of value flow. Success is achieved by transforming manual, high-friction processes into automated, "fire and forget" workflows developed through continuous discovery and a deep focus on the developer as the primary customer.
Think of a modern platform as a well-maintained highway; while developers still drive the vehicle of business logic, the platform team removes the toll booths and clears the debris, allowing them to reach their destination at maximum speed without having to navigate the rough terrain of infrastructure on their own.

📋 Key Points

  • Adopt a Platform as a Product mindset, focusing on the developer as the primary customer and optimizing for flow rather than just managing infrastructure.
  • Prioritize the reduction of developer cognitive load, enabling teams to dedicate more mental energy to business logic instead of low-level platform details.
  • Transition from a traditional ticket-oriented service model to a modern product-oriented approach that measures success through speed and platform adoption.
  • Use continuous discovery techniques—such as shadowing, interviews, and surveys—to align platform development with actual user needs.
  • Leverage instrumentation to identify critical paths and bottlenecks before attempting to optimize processes.
  • Implement changes through Small Safe Steps (3S) and parallel changes to deliver value quickly while minimizing risk.
  • Transform manual, error-prone release processes into automated "fire and forget" workflows (e.g., via Slack commands) that provide fast, clear feedback.
  • Maintain the Thinnest Viable Platform (TVP) by building only what is essential and leveraging external services or open-source software to avoid over-engineering.
  • Ensure adoption is not forced, but rather earned by creating tools that are easy to use and demonstrably reduce friction for stream-aligned teams.
  • Evolve organizational structures using Team Topologies to facilitate scalability and maintain clear ownership across different domains.
  • Promote engineering excellence by modeling best practices such as trunk-based development, TDD, and continuous delivery within the platform team.