Blake Hallisey

I turn messy AI problems into products that ship.

Senior product manager. I build agents and ship them at Sephora, Dell and Google.

An AI agent that answers from my work.

What I've shipped

The short version. Or just ask the agent.

One of Dell's early ML-powered platforms

Wired NLP into the CRM so RFPs went from weeks to hours, cutting turnaround 85% and recovering $15M-$20M in annual sales.

47-skill agentic stack, Sephora

A multi-agent system for sprint planning, dependency tracking, and status reporting inside regulated loyalty fintech.

Two-key encryption, Dell PowerMax

End-to-end encryption under HIPAA/PCI/SOC2: 4:1+ compression, $30M+ in at-risk revenue, zero audit findings.

For fun: Squid AI · First Atom

An enterprise-agent blueprint tool and a biotech site built under a no-fake-science rule.

The work

The work01 / 08
The idea

Anyone can build an agent. Knowing it's right is the job.

How I think

I was a software engineer before I was a product manager, so I start from what the system can actually do. Since then I have worked across cloud, enterprise infrastructure, security and AI, mostly turning a vague technical problem into something a team can ship. I specify the evals, ship the agents, and own the call on what is ready.

How I thinkLayer 01 / 05

Let's build.

Ask the agent, or take the resume with you.