Ethan Schreiber
Product · Platforms · APIs
Austin, TX
Senior Technical Product Manager Platforms · APIs · Integrations Austin, TX

Ethan Schreiber

Product manager for internal platforms — the APIs, integrations, and tools employees rely on to serve members.

§ 01
I work on internal platforms — CRMs, APIs, integrations, and the tools employees use every day.

Currently at a top-10 credit union, where I lead API decoupling, an effort to recover more than 20,000 hours of employee time lost to clunky customer-service tools, and the APIs that let employees take action for members — replace a card, close an account, resolve a dispute — across our card, banking, and digital vendors.

Outside of work I build AI agent systems. The longest-running was Choicescout, a content pipeline that used Claude for research and writing and Gemini for images.

Ethan in a wide grass valley with mountains behind him. — Lamar Valley · Yellowstone —
§ 02

Selected work

A few pieces from current and recent roles — see more.

First Tech FCU API Strategy

Decoupling legacy APIs

Identified which capabilities were safe to expose externally versus internal-only, then set the versioning and rollout sequence so internal tools could ship independently of member-facing surfaces.

First Tech FCU Operational Efficiency

20,000 hours of internal time recovered

Traced the loss to tool architecture: manual handoffs, redundant approvals, and missing integrations. Led the redesign — automating handoffs and connecting card processing and core banking directly — so employees spend less time fighting their tools.

First Tech FCU AI · Agents

Internal AI agent, 1,000+ daily questions

Product vision, prompt engineering, and rollout for an internal AI agent (Pega Knowledge Buddy) used by customer service reps. Built the business case, made the build-vs-buy call, and ran the rollout. Saves roughly 1,200 hours a year.

Personal Multi-agent

Choicescout

An AI content system I built and ran. Specialized agents researched, wrote, and published affiliate product reviews end to end, with conventional code enforcing the quality checks the AI couldn't be trusted with on its own.