Abe Reyes
Founder, Ascendant Ventures — Building Foundary
I build systems that make complex software delivery reliable, auditable, and safe to operate at enterprise scale. Two decades at the intersection of platform architecture, reliability engineering, and organizational governance.
Now
Founder, Ascendant Ventures. Building Foundary — governed agentic delivery for enterprise teams.
Previously
VP / Principal Architect at Salesforce. Chair, C360 Availability Council. CTO of MapAnything through acquisition.
Expertise
Platform architecture, reliability & SLOs, delivery governance, agentic systems, enterprise-scale operations.
Two decades building systems that hold.
From startup CTO to enterprise VP, the through-line has always been the same: make delivery reliable enough to trust.
Founder
Building Foundary — a platform for governed agentic delivery. Designed for enterprise teams that need AI-assisted development to be as auditable and policy-bound as the rest of their software process.
VP / Principal Architect
Led enterprise architecture for CRM (Sales, Service, Field Service). Chaired the C360 Availability Council, aligning cross-cloud reliability objectives, incident standards, and scalable operating practices across one of the world's largest SaaS platforms.
CTO
Owned technology strategy and execution from growth stage through acquisition. Drove cloud-native platform delivery and operational maturity that positioned the company for a successful Salesforce integration.
"AI-assisted development is only credible
when it is governed."
Every reliability and governance problem I've seen at scale comes down to the same thing: process that works in demos but breaks under production pressure. Foundary exists to make agentic delivery structurally trustworthy — not just fast.
Approvals are durable
Every gate is recorded and verifiable. No decision is ephemeral, no approval can be retroactively disputed.
Runs are auditable
Every delivery run produces a sealed record. What happened is knowable — always — without relying on agent self-reporting.
Execution is policy-bound
Agents operate within defined constraints. Policy is structural, not advisory. The same gates apply to every run, every time.
Let's talk.
If your organization is thinking seriously about governed agentic delivery — or if you're evaluating how AI fits into your software process — I'd welcome the conversation.