RGDSolutions
← Back to selected work

LaunchDarkly · Experimentation · AI product engineering

Experimentation and AI, ready for rollout.

Ricardo contributed to two areas of LaunchDarkly’s platform: Experimentation and AI Configs. His work combined frontend development for experimentation workflows with architecture, APIs, and technical leadership for a new AI product, helping carry AI Configs from inception through beta to General Access.

Ricardo’s work as frontend engineer, core engineer, and tech lead

Experimentation: make complex workflows usable

As a frontend engineer on the Experimentation Platform, Ricardo developed complex React applications with reusable components, dynamic data handling, and responsive layouts. The project focused on improving feature experimentation workflows and collaboration for engineering teams. His stack included TypeScript, React, Redux, Storybook, and REST API integration.

From experiment design to results

The screenshots show the product workflows: defining a hypothesis and funnel metrics, allocating traffic across variations, reviewing experiment results, and connecting feature-flag targeting with experiment holdouts. They illustrate the application context for Ricardo’s frontend work; the example conversion figures are product demonstration data, not business outcomes attributed to his contribution.

AI Configs: from inception to General Access

As a core engineer on AI Configs, Ricardo contributed to the architecture, APIs, and data model for a new addition to LaunchDarkly’s flag-management portfolio. He worked with Product, Design, Security, and SRE on reliability and auditability, helping drive the product from beta to General Access.

Approvals and evaluation before rollout

Ricardo served as tech lead for AI Configs approval workflows, including role-aware rules, audit logs, and guarded change management. He also led Test Runs, a sandbox and evaluation framework for safely executing and comparing configuration changes before rollout.

Team leadership and product impact

Ricardo managed three team members, mentored new engineers, and addressed technical debt that slowed developer productivity. During this work, AI Configs reached the company’s first-quarter revenue goals two months into the quarter—a product-level milestone reported by Ricardo, achieved through the broader team’s work.

Inside the product

LaunchDarkly experiment results with a hypothesis summary, exposure chart, and variation comparison table.
Review experiment results and compare variations
LaunchDarkly funnel experiment preview showing metrics, audience targeting, and traffic allocation across three variations.
Define funnel metrics and allocate experiment traffic
LaunchDarkly feature-flag targeting with a prerequisite holdout flag and an experiment traffic allocation rule.
Connect feature targeting with experiment holdouts

LaunchDarkly Experimentation · Screenshots supplied by Ricardo · Ricardo’s project on A.Team

Building something similar?

Tell us where you are today and what you need to ship next.

Discuss your project