Onboarding Docs: SageMaker and AWS-Wide Improvements

Documentation improvements that went beyond a single product—identifying systemic friction points and driving changes that impacted documentation across all of AWS.

Scope: AWS-wide Impact
Collaboration: SageMaker AI, AWS IAM Docs Team
Focus: Onboarding Documentation

View the Live Documentation

The Problem

From my first days at Amazon SageMaker AI, a consistent theme emerged: Not only me, but the team, PMs, and engineers all struggled with the onboarding documentation. Getting started with SageMaker AI meant navigating through complex IAM (Identity and Access Management) setup, domain setup, and service prerequisites—and the existing documentation wasn't making this journey smooth.

Rather than accepting this as just how it is, I decided to systematically investigate and address the root causes.

My Approach

Phase 1: Deep Dive into User Journeys

  • Identified and documented all personas using SageMaker AI—from individual data scientists to enterprise administrators
  • Walked through each persona's workflow step-by-step, taking careful notes on pain points, confusion, and gaps
  • Created documentation improvements for SageMaker AI-specific onboarding paths

Phase 2: Identifying Systemic Issues

While improving SageMaker AI documentation, I discovered the friction wasn't just in our docs—it was rooted in how the AWS IAM (a foundational service used by all AWS products) onboarding documentation handled certain onboarding scenarios. The IAM documentation is deployed AWS-wide through an entity, so by identifying the friction points there I was identifying them across all onboarding docs for AWS services, not just SageMaker AI.

Phase 3: Cross-Team Collaboration

  • Documented my findings and proposed solutions with clear persona paths
  • Reached out to the AWS IAM team to share my analysis
  • Collaborated with IAM to update the underlying entity and documentation
  • Changes were adopted and rolled out across AWS documentation

Impact

The IAM improvements I identified and helped implement impacted documentation and user experience across all of AWS—not just SageMaker AI. This improves the entire platform, not just one service.

What This Work Demonstrates

  • Systems thinking: Looking beyond immediate symptoms to identify root causes, even when they cross team boundaries
  • Initiative: Taking ownership of problems that "everyone knows about" but no one has tackled
  • Cross-functional influence: Building relationships and driving change across organizational boundaries without direct authority
  • User advocacy: Prioritizing actual user experience over internal convenience

Documentation Structure (SageMaker AI Setting Up)

The improved SageMaker AI onboarding documentation I developed:

  • Setting up SageMaker AI
    • Complete Amazon SageMaker AI prerequisites
    • Use quick setup — Streamlined path for simple use cases
    • Use custom setup — Detailed path for enterprise requirements
    • Domain overview
      • SageMaker AI domain entities
      • Choose an Amazon VPC
    • Supported Regions and Quotas

Unpublished Work

A larger revamp of the "Setting up SageMaker AI" documentation was completed and reviewed with the PM, but with the sudden departure I was unable to publish. The improvements described above represent just the beginning of that larger effort.

Broader Context

This work exemplifies my approach to technical writing: documentation isn't just about describing features—it's about understanding user journeys and removing obstacles, wherever they exist. Sometimes the most valuable documentation work happens at the boundaries between teams and products.