ABOUT · MAST LABS
In March 2008, I wrote down what I wanted to build.

I called it PharmsNet. The ability for pharmacies to digitally track prescribed medications across a shared network. Patients, doctors, and pharmacies connected. Prescriptions documented on a common digital board. Pharmacists notifying doctors of conflicts. Session timeouts. Encrypted records. Medication data masked except in emergencies, accessible only to authorized users.

I had the architecture. I had the timeline. I had the business plan. I had a meeting scheduled for April 9th with an engineer, kept secret until we filed. The person I wanted to build it with didn't want to.

PharmsNet never got built.

The sell line I wrote in that notebook: Being in the loop will save lives.

I was 26.

2008
PharmsNet · March 16, 2008
Concept · Architecture · Business Plan
Never built
THE THROUGHLINE
I spent the next seventeen years becoming capable of building it.

Not in a straight line. Not with a plan. Through the work.

Eat In Chef taught me that technology only works if the surrounding structure can absorb it. I built the matching logic, the onboarding system, the operational layer. I was bartending and serving tables to cover rent while running the company and leading a $3M seed round. The technical co-founder left. I kept building. Four cities. Thirty-six chefs. Three thousand on the waitlist.

I stopped thinking in features and started thinking in flows.

At SiriusXM and Pandora, the variables changed. Two hundred million assets. Seventy million users. A taxonomy that was constraining what the intelligence layer could do. I rebuilt it. Replacing a third-party contextualization system blind. Designing the first structured classification framework. Collaborating with scientists to redefine the schema after a major acquisition brought in a genomic model that was constrained and outdated.

What truly excited me was the data. Modeling on whiteboards. Defining attributes, relationships, clusters, interactions. Mapping scenarios to shape outcomes.

That's where the brains thesis started. Not from a framework. From the work.

THE CONVERGENCE
Then came a regulated digital pharmacy. Everything converged.

Healthcare operations. PHI constraints. Fragmented teams. Ambition without mature infrastructure.

I built the connective tissue. Orchestration layers. Communication backbone across SMS, voice, chat, and telephony. AI routing systems. Governance patterns. Not prototypes. Production systems. The kind that run 24/7 and carry clinical accountability.

The system I built was a direct descendant of what I wrote in that notebook in 2008. Patient records tracked through a network. Pharmacist-to-doctor notification. Prescription conflicts flagged. Confidential, encrypted, session-controlled.

The intelligence layer is always the moat. And it almost always belongs to no one.

Mast Labs is that layer. Seventeen years in the making. mastOS is how we deploy it.

2008
PharmsNet
The idea. Never built.
2014
Eat In Chef
First lesson in data and behavior mapping.
2018
SiriusXM · Pandora
Intelligence layer at scale. 200M+ assets.
2023
Mast Labs
The platform, building.
NOW
In production.
Agentic framework. Azure-native.
Architecture proven at scale. mastOS is the build.
FOUNDER
FOUNDER & CTO
Ara Galarza
Founded Eat In Chef in 2014. Built and scaled to four cities, thirty-six chefs, and three thousand on the waitlist. Bartending, serving tables, and cooking privately to cover rent while running the company and leading a $3M seed round. Built the board. Carried it.
At SiriusXM and Pandora, rebuilt the genomic taxonomy and content intelligence system underpinning recommendation, personalization, and content discovery across a 70M+ user platform. 200M+ assets reclassified. Foundation for $215M+ in annualized revenue impact.
At a regulated digital pharmacy, led AI platforms from zero to production. PHI-safe multi-agent orchestration across patient acquisition, triage, lifecycle, and renewal. 500K+ autonomous interactions. 96% resolution. Sub-1% error rate.
Mast Labs was built during recovery from a severe accident in 2024. The work didn't stop. Neither did the building.
The perspective is different because it was earned the expensive way.
“The intelligence layer is always the moat. And it almost always belongs to no one.”
Ara Galarza · Mast Labs
WORK WITH US
Mast Labs takes a small number of engagements per quarter.
Fit matters more than fee.
Start a conversation