For founders, product teams, and investors
I build practical, production AI systems.
I help you decide what to build, fix, release, or buy, then do the technical work those decisions call for. After 30 years architecting software, I build and run AI systems in production myself.
AI App Review: a focused review capped at $600, or two full days in your code for $1,500. For anything bigger, the first conversation is free.
- 30 years building enterprise software
- 18 years as Technology Architect at Old Republic Home Protection
- 10 live products I build and run, including Klass Personal OS
- 45 posts on production AI engineering
Start here
AI App Review: an outside read of code your team built with AI coding tools
AI coding tools write what you asked for and stop there. The checks a careful engineer adds by habit, like authorization on every record, server-side validation, and limits that hold on every path into the data, are the parts they skip and a demo never exercises.
$1,500, one time
- Two full days reading and running your code by hand
- A written report ranked by what to fix now, soon, or leave
- A 90-minute walkthrough with your team
- Fixes quoted separately, only if you want them
Reviewing an AI build? Start with the free kit.
Get the questions I ask before release, my ship pipeline, a real review example, plain-English terms, and editable project files. It's built for founders and people using AI coding tools to make software.
Get the AI Build Review KitOther ways to work together
Start with the review or a fixed-scope assessment, bring me in to build, or retain me as your ongoing technical leader. Every engagement is hands-on - you work directly with me.
Focused Code Review
Building your own app with AI tools? I can answer one question about the code in a short written review, billed only for time used and capped at four hours.
AI App Review
Two full days reading and running code your team built with AI coding tools, then a ranked report of what to fix first and a 90-minute walkthrough.
Architecture & AI Assessment
A focused look at your system or AI stack: what it costs to run, where the real risk is, how to right-size the spend, and your realistic options - written up as a prioritized plan.
Technical Due Diligence
For acquisitions and investments. I assess the codebase, infrastructure, and team, then deliver a written report with a realistic post-acquisition picture.
AI Build & Implementation
Custom AI built to production standards - agents, integrations, grounded chat and search over your own data, and automation, with guardrails and cost controls from day one.
Fractional CTO Retainer
Ongoing technical leadership for companies that need a senior technical voice without a full-time hire. Up to 15 hours per week.
Case Study
Automating a matchmaking firm's intake and consultation pipeline
A boutique matchmaking business was collecting client intake by hand and re-keying it, along with notes from every consultation call, into their matchmaking CRM. I built the pipeline that removed the manual step.
What I built
- A custom Angular intake form on a Spring Boot backend, with draft autosave, field-level revision history, and validation
- On submit, an asynchronous push of the intake data into their matchmaking CRM (SmartMatchApp) over its REST API, with an email notification linking straight to the newly created client profile
- A Zoom integration that captures AI-generated consultation summaries via webhooks and a Server-to-Server OAuth app, then parses and feeds them into the same CRM
- Business email migration, hosting, domain cutover, and analytics and privacy setup to take it all live
The outcome
Client intake and consultation notes now flow into the matchmaking CRM automatically instead of being typed in twice by hand. Launched in production in April 2026.
Systems I build and run
A few of the AI systems I run in production every day. Not slideware - working software I maintain myself.
Personal AI Agent Infrastructure
A self-hosted MCP server that gives Claude live access to my real systems - a 59-tool server spanning 12 functional domains (email, calendar, files, CRM, git, storage, and more), running scheduled agents in a single production deployment.
Autonomous AI Operations System
A fleet of Claude agents that runs continuously against an organized knowledge base - handling email triage, CRM sync, content publishing, calendar management, and scheduled workflows on crons from every 10 seconds to weekly.
Remote MCP Servers for AI Connectors
Two live products expose their own remote MCP server - 73 tools and 21 tools - that Claude and ChatGPT connect to over the web, secured by an OAuth server I wrote so each connector is scoped to a single tenant.
Latest writing
Notes from systems I actually run: the failure, the cause, and what I changed.
Most findings in a technical review get priced in. A red flag is different: a gap between what the product promises and what the code actually does. Examples from reviews I've done this year, including my own product.
A doc comment said the function spliced bytes into the original file. It parsed the file and rebuilt it, silently dropping anything the parser didn't model, on every single call.
A set of tools was excluded from what the model gets offered, with a doc comment explaining the hazard. Nothing stopped the model from calling them anyway.
Why me
I've been building enterprise software for 30 years, 18 of them as the Technology Architect at Old Republic Home Protection, where I stewarded a 20-year-old legacy platform, built its modern replacement, and then designed a second greenfield system on Azure. I still write code every day and run AI in production, so when I tell you a project is harder than it looks, it's because I've done the work recently myself.
Questions I get asked
Is AI actually the right tool for my problem?
Sometimes the honest answer is no, and I'll tell you when it is. A lot of what gets sold as "AI" is a script or a better process. Where AI genuinely fits, I'll show you where it pays off and where it will quietly waste money - and I build it with cost controls and guardrails so it stays reliable.
How is this different from hiring a consultant?
A consultant typically delivers a report and leaves. A fractional CTO is embedded in your company - attending meetings, working with your team, making decisions, and being accountable for outcomes. The relationship is ongoing, not transactional.
Do you work with companies that don't have an engineering team yet?
Yes. In that case the work often focuses on helping you decide what to build vs. buy, choosing a technology stack, defining what your first engineering hire should look like, and evaluating vendors or development agencies. Getting the foundation right before you build on it matters more than most founders realize.
How many clients do you take on at once?
I limit retainer clients to three at a time. I won't take on a fourth client if it means the first three get diluted attention. If I'm at capacity when you reach out, I'll tell you and we can talk about whether a project engagement makes sense in the meantime.
What if I only need a few hours of advice?
The Architecture & AI Assessment is the best fit for a bounded engagement. If you have a specific question that doesn't fit a defined scope - a vendor decision, a second opinion on a technical proposal, a hire evaluation - reach out and describe it. I'll tell you whether I can help and what that would look like.
Ready to talk?
The first conversation is free. Tell me what you're dealing with and I'll tell you honestly whether I can help.
Book a 30-minute callOr reach me directly: info@coreyklass.com · 925-705-5083