Alaska Airlines — Custom Staffbase Engineering for 30K+ Employees

Overview

Alaska Air Group operates a large internal employee platform on Staffbase used across frontline and corporate teams. The platform supports internal communications, operational updates, content discovery, and employee engagement workflows in an environment where reliability and clarity matter.

I worked through Favish, in collaboration with LocalWisdom, to build and maintain custom production-grade Staffbase plugins and widgets running inside Team AAG, the employee platform used across the organization.

Scope of engagement

End-to-end engineering across multiple Staffbase-based products, including enterprise search, bulletin acknowledgement flows, alerts, employee recognition, gamified training, and shared UI infrastructure.

What I owned

• Owned delivery of multiple custom Staffbase widgets and plugins running in production on Alaska Air Group's employee platform.
• Built a custom Elasticsearch-backed search experience for internal content discovery across the platform.
• Implemented acknowledgement and bulletin workflows for employee communications with action-oriented UX.
• Engineered key parts of a native peer-to-peer recognition product replacing a legacy PowerApps implementation.
• Solved platform-specific mobile navigation and deep-linking issues inside the Staffbase app environment.
• Managed API integration changes, migration work, and production fixes across evolving Staffbase platform dependencies.
• Contributed shared React and TypeScript infrastructure reused across multiple production widgets.
• Handled live production debugging and reliability work across frontend, integration, and platform layers.

Selected systems

• Smart Search — Custom Elasticsearch-backed search widget for internal content discovery, with platform-aware filtering and mobile deep-linking support.
• Shoutouts Plugin — Native peer-to-peer recognition product with submission flows, feed, history views, push and email notifications, and PeopleSoft-backed recipient selection.
• Bulletin Acknowledgement Widget — Action-oriented communications widget helping employees surface, review, and acknowledge required internal updates.

Enterprise search and information access

One of the core workstreams was a custom search experience designed to improve how employees find internal pages, articles, and other platform content. The widget used Elasticsearch-backed search flows combined with Staffbase APIs to surface relevant results while respecting platform visibility rules and content access constraints.

This work included search UX, integration behavior, reliability improvements, and handling platform-specific edge cases, particularly around mobile navigation and deep linking inside the Staffbase app environment.

Employee communications workflows

I also worked on communication-focused widgets designed to make important internal updates more actionable. This included bulletin acknowledgement flows, content filtering, multi-channel support, and UX improvements intended to reduce friction for employees consuming required communications.

In practice, this meant building interfaces that helped teams see what required action, what had already been reviewed, and what information still needed attention, all within the company's core employee platform.

Recognition and engagement systems

A major initiative was a native Staffbase recognition plugin designed to replace a legacy PowerApps-based workflow. The system included shoutout submission flows, feed experiences, history views, email and push notification paths, and directory-driven recipient selection backed by Staffbase and PeopleSoft user data.

I also contributed to a gamified training experience tied to an internal loyalty-program rollout, including leaderboard behavior, user-data handling, and recurring scoring-related workflows.

Technical complexity

The engineering challenge was not just building UI. The work required operating inside Staffbase's plugin model, integrating with multiple APIs and identity flows, handling mobile WebView constraints, working through platform changes, and shipping functionality that remained stable under real production usage.

Several systems depended on careful handling of authentication, tenant-aware routing, session-sensitive API behavior, external data integration, and long-tail edge cases that only appear in live enterprise environments.

Selected engineering wins

• Shipped production AI content moderation (Gemini) for the recognition product, designed fail-open so an AI outage never blocks an employee: retry with exponential backoff, and a 42-case evaluation suite that runs concurrently in ~14 seconds and gates every model change.
• Made a 64,000-user directory sync complete reliably after it failed on every run: replaced offset pagination with keyset pagination, batched large upserts, and added deadlock-aware retries — the same crawl now scales toward millions of users.
• Migrated notification dispatch from an in-process fire-and-forget pattern to Google Cloud Tasks with authenticated, at-least-once delivery and a reconciliation job — eliminating notification loss on pod restarts.
• Drove the security-first authentication design from kickoff: evaluated widget, plugin, and hybrid auth models with explicit risk framing, and resolved a mobile OAuth blocker in direct collaboration with Staffbase engineering.
• Fixed an unreproducible iOS tap-swallowing bug affecting airline crews by bypassing framework event delegation with direct DOM listeners — labeled transparently as a blind fix, validated by the client the same day.

Platform scale

This work ran inside an employee platform used across a large airline group, supporting 30,000+ employees across frontline and corporate functions. The broader platform achieved strong adoption and engagement, which raised the quality bar for reliability, usability, and operational correctness.

Outcome

The result was a set of custom internal products that improved content discoverability, communication workflows, employee engagement, and day-to-day platform usefulness in production. The work helped extend the capabilities of the employee platform beyond the base Staffbase product through tailored, business-specific software.

From a delivery perspective, this was high-trust product engineering in a live enterprise environment: building systems that had to work reliably for a large distributed workforce, across desktop and mobile, with minimal room for failure.