About Mike Hostetler

Builder first. Always have been.

I wrote my first lines of code in QBasic as a kid, and I’ve been shipping software ever since. Over three decades later, the tools have changed (from PHP and Drupal to jQuery to Elixir to AI agents) but the drive is the same: I care less about writing elegant code and more about building things that work and matter.

My career has been a series of bets on emerging technology. In 2009, I co-founded appendTo with fellow contributors to the jQuery open-source project. We bootstrapped that company to 30+ people and millions in revenue, delivering training and consulting to Fortune 500 organizations like Microsoft, Time.com, and ETrade. I was part of the jQuery Core team and the jQuery Foundation from its founding era, and I co-authored O’Reilly’s jQuery Cookbook* along the way.

From there I moved into enterprise leadership. At Cars.com, I was part of the leadership team that managed 120+ engineers across a $650M revenue platform, helped execute a full technology re-platform, transitioned the company to remote work during COVID, and drove $3M in annual vendor savings. That experience (operating at scale, managing complexity, and translating between engineering and the C-suite) shaped how I think about technology leadership today.

I’ve also spent years in the blockchain and crypto space as COO of Versatus.io, where I led strategy, operations, and infrastructure for a blockchain startup. I built projects across TypeScript, Elixir, and Rust, contributed to the Solana ecosystem, and learned firsthand what it takes to build in a space where the ground is constantly shifting under your feet.

What I’m building now

Today, my primary focus is Jido (jido.run), an open-source autonomous agent framework built on Elixir. The underlying runtime was originally designed for telecom systems that never go down, which makes it ideal for running thousands of AI agents simultaneously with built-in resilience. Jido can power 1,500+ concurrent agents on a single machine, and the ecosystem now spans AI model integration, browser automation, and a unified client library supporting 665+ models across 11 providers.

My thesis is simple: the bottleneck isn’t making a single AI agent smarter. It’s making them work together. Agent orchestration is the hard problem, and most of the industry is solving it with duct tape. There’s a concept called “Intelligence Hyperinflation” that captures what happens when organizations deploy disconnected AI bots across departments. You don’t get intelligence, you get expensive noise. The fix isn’t better prompts. It’s better architecture. At every organization I’ve been part of, I’ve pushed for AI orchestration at the enterprise level, not just point solutions.

I also serve as Director of Engineering at ServiceCore, where I’m building an AI-first engineering culture and scaling a senior technical team.

How I lead

Building great technology means building great teams. I’m a practitioner and coach of the Extreme Ownership philosophy: every outcome, good or bad, starts with leadership taking full responsibility. I didn’t always lead this way. Earlier in my career I was heads-down on the technical work, and it took great mentors, hard lessons, and a lot of experience to understand that taking care of your team is the job. Over time, through a focus on servant leadership and a commitment to owning every outcome, I’ve grown into a leader who embodies these principles daily.

Coaching is core to how I operate. I’ve designed custom engineering coaching programs, mentored founders navigating their first technical hires, and delivered leadership training to organizations of every size. The common thread is ownership: of the mission, of the mistakes, and of the growth that comes from both.

How I think

I bring an unusual combination to the table: deep hands-on engineering and formal business training. I earned my Executive MBA from Northwestern’s Kellogg School of Management, focused on venture capital and acquisition entrepreneurship, on top of a B.S. in Computer Science from UW-La Crosse. I’m as comfortable deep in code as I am in a board room.

A few beliefs that guide my work:

Agents need solid foundations before you add AI. There are 40+ years of agent research in computer science. We shouldn’t throw it all out just because the latest AI models are exciting.
You cannot automate chaos. Fix your processes before you apply AI to them. A highly efficient bad process is still a bad process.
Software engineering is thriving; rote software development is dying. The mechanical work of writing code is being automated. The systems thinking, orchestration, and problem definition: that’s where the real work lives.
Ship, then iterate. I’ve always been a builder who learns by doing, not by theorizing.

Let’s connect

I build in public, write about AI agents and engineering leadership, and I’m always open to conversations with founders, engineers, and operators who are building the next wave.