Restoring Predictability to Platform Delivery

Re-architecting core developer workflows to support stable delivery leading up to GA.

The work unfolded in an environment where system capabilities were still evolving—requiring continuous translation between technical constraints and meaningful user interactions.

Role: UX Manager Domain: Enterprise Engineering Platform Timeline: 12-week full redesign

On this page

Flow diagram showing the end-to-end CI/CD pipeline — from GitHub through Jules CI, Spinnaker CD, and into DEV, QA, STAGE, and PROD environments, with Release Management and Release Orchestration layers

This initiative required more than interface updates—it meant restoring structural clarity to a delivery system in flux.

Mindset Reframed the problem from a feature migration effort to a delivery integrity challenge — positioning UX discovery as the structural front end of the development process.
Scope Moved beyond individual screens to workflow architecture, clarifying how engineers navigate promotion, diagnostics, reporting, and pipeline states across the SDLC.
Process Established a discovery-first sequencing model that aligned design exploration with engineering planning cycles, improving estimation accuracy and reducing downstream churn.
Delivery Identified that UX was entering the process too late to meaningfully influence delivery and restructured the workflow to integrate UX earlier, enabling better planning, anticipation of engineering needs, and more reliable delivery under pressure.

Context

Platform in Transition Under Time Pressure

When I stepped in as Experience Lead for Build & Release, the new Pipeline workflows in IEP were the next major capability scheduled for General Availability.

The platform had already launched in beta, but several core engineering workflows remained fragmented across legacy patterns and evolving backend services. Pipelines functionality also existed in a separate system (JET), but architectural differences and workflow inconsistencies made direct migration structurally unsound.

At the same time, developer adoption depended on bringing critical capabilities into the new platform before GA. The challenge was not simply migrating features—it was establishing a coherent workflow architecture capable of supporting predictable delivery across the engineering lifecycle.

Without structural clarity, migration would have amplified the existing friction.

GA Timeline Pressure

The platform needed to integrate essential developer workflows within a constrained delivery window while backend development was still actively evolving.

Cross-System Migration Risk

Existing capabilities lived in legacy systems with different architectural assumptions, making lift-and-shift migration impractical.

Fragmented Workflow States

Promotion flows, diagnostics, reporting, and pipeline states had grown organically, producing fragmented mental models for engineers.

Reactive Delivery Patterns

Historically UX entered the process late, limiting its ability to shape architectural clarity and forcing design decisions to respond to engineering constraints rather than guide them.

Lucidspark board showing 7 parallel systems in the Build and Release Ecosystem before intervention — Evidence Widget, Deployment App, Release Manager, Runway, Earhart, Release Coordinator, Pipelines, and Policies with no shared structure — Build & Release Ecosystem, pre-intervention — 7 parallel systems, no shared structure

Strategy

Designing for Predictable Delivery

Rather than relying on existing UI models, this work required defining interaction patterns that aligned system behavior with how engineers reason about delivery, failure, and recovery. This led to three core principles that guided the design:

Clarifying Workflow Structure Explicit

Mapped developer journeys across pipeline states before designing UI artifacts. This exposed structural inconsistencies between systems and clarified the workflow architecture required to support a coherent developer experience. Early discovery surfaced points of friction that informed how the system should organize and present work across stages.

Enabling Predictable Delivery

Introduced a discovery-first execution model that embedded UX exploration into the delivery cycle rather than treating it as a separate phase. By intentionally sequencing research, design, and engineering planning, the team established a more predictable path from discovery to implementation—reducing ambiguity and improving alignment across stakeholders.

Scaling Through Leadership Multiplication

With a small team supporting a complex platform, I established operating practices that allowed the group to function as a coordinated unit. By balancing hands-on architectural definition with enabling the team to own execution, the work scaled more consistently—maintaining coherence across workflows as delivery expanded.

End-to-end workflow architecture map showing all pipeline modules — Grafana, Pipelines Widget, Evidence Widget, Policy Service, Deployment Widget, Release Manager, and Release Orchestrator — connected across the full SDLC — Pipelines Executions — end-to-end workflow architecture across the full SDLC, connecting all 8 modules

Execution

From Fragmentation to Structured Workflows

Within a twelve-week window, my team and I delivered eight core workflow modules, integrating legacy functionality into the new platform while establishing clearer developer journeys. Rather than designing isolated features, the work focused on restructuring how engineers move through pipeline states and supporting functions across the SDLC. This required not just reorganizing workflows, but defining how they should be experienced across stages.

Sequencing the Work Intentionally

Previously disconnected promotion flows were re-architected into structured stages with transparent authorization logic and integrated validation states—aligning system behavior with engineers' mental models. Work was delivered through a sequenced sprint model:

Sprint 1

Discovery & backend logic alignment

Mapping system states, permissions, and workflow transitions with engineering teams.

Sprint 2

Dev-ready experience design

Delivering interaction models and UI architecture aligned with implementation constraints.

This structure allowed design exploration and technical feasibility to evolve together rather than sequentially.

Key Modules Delivered

Evidence & Promotion

Structured promotion workflows and validation states supporting controlled progression across pipeline environments.

Evidence and Promotion module — UI showing evidence upload, code quality checks, and promotion status

Pipelines Overview

Improved visibility into pipeline health and developer actions required to progress builds.

Diagnostics

Unified troubleshooting surfaces across logs, timelines, and policy checks.

Reporting

Structured reporting to support auditing and operational oversight.

Sprint sequencing plan showing 6 sprints across November 2024 to January 2025, with modules moving through discovery, Figma specs, validation, and UX delivery phases — GA delivery plan — 6-sprint sequencing across Evidence, Listing, Diagnostics, Waterfall / Logs, and Reporting modules, November 2024 – January 2025

Impact

Delivery Integrity Restored

The intervention stabilized delivery sequencing and demonstrated the value of structured UX leadership within a high-pressure enterprise environment.

6 Sprints Structured

Intentional sequencing of research, design, and engineering planning across a 12-week delivery window.

100% Planned Work Delivered

All committed engineering work completed within the execution window.

8 Core Workflows Re-architected

Fragmented pipeline states consolidated into coherent journeys across the SDLC.

45% of Legacy Platform Traffic Transitioned

The redesigned workflows represented nearly half of legacy system activity and were critical to achieving GA readiness.

These results reflected more than delivery performance—they demonstrated how intentional sequencing of discovery, design, and engineering planning restored predictability to a critical developer platform—clarifying how the system behaved and increasing user confidence.

Reflection

What I Learned

This experience reinforced that designing at system scale requires more than aligning teams or refining interfaces—it requires shaping how complexity is made understandable.

As systems grow more interconnected and less deterministic, clarity becomes a design responsibility. It’s not only about structuring workflows, but about helping people anticipate outcomes, understand system behavior, and make informed decisions within it.

I also learned that predictability doesn’t come from control alone, but from thoughtful sequencing, shared understanding, and continuous alignment across disciplines. These conditions allow complex systems to become not only usable, but trustworthy.

← All Work