I turn business knowledge, proprietary methods, and expert judgment into AI systems that improve revenue, delivery, and decision-making.
For more than 20 years I have built and scaled platforms while their markets were still forming.
I led Waze LATAM from launch to 15M users across six markets with zero paid acquisition, scaled Twitter, Spotify, LinkedIn, and Snapchat across emerging markets, and co-founded a personal AI platform where, as COO, I led product strategy and GTM to 140K daily active users.
I now build AI to run on your own knowledge, rules, and judgment.
Featured in Bitcoin.com, HackerNoon, Unite.AI, TechTimes, Blockworks, and The Defiant.
Every project starts with a business outcome: keeping every client and upsell opportunity on track, reducing four hours of daily investment research to under two minutes, or applying a proprietary method consistently across hundreds of cases.
Together, we identify where AI can create the greatest value. I design the solution, build it around your knowledge, data, rules, and judgment, and implement it in Claude, ChatGPT or any LLM of your choice.
The result is a working system you own and operate.
We map the work, sources, rules, repeat points, and review burden.
Simple workflows can usually move straight into a fixed build. Larger systems start with a focused first phase.
You receive the workspace or workflow, the training to run it, and the correction step that keeps it accurate as the business changes. You own it and run it after that.
Your methodology defines how you invest. I turn it into a system that applies those rules across research, portfolio monitoring, watchlists, limits, and position context. It does not decide for you. It enforces the discipline you defined. See how it works →
The client history, notes, reports, and follow-up logic of an expert practice, in one workspace the work runs from.
Team workspaces that bring repeated output closer on the first pass, so review time falls without losing senior judgment.
Campaigns, announcements, and approval prep built around founder judgment, approved claims, campaign history, and roadmap context.
A private investor's proprietary methodology now runs across daily research, portfolio monitoring, opportunity discovery, and analysis of approximately 300 securities.
I built her rules, source requirements, validation standards, limits, and exceptions into the system. A process that previously took four hours each day now takes under two minutes.
Read the case study →I built an AI CMO that applies the same positioning, offers, audience knowledge, voice, standards, and strategic decisions across the business.
Built on more than 20 years of marketing experience, the system has created a website, produces short- and long-form content across multiple platforms, creates visuals, and develops frameworks, courses, campaigns, and other business assets.
New decisions and corrections are incorporated into the system, so every workflow draws from the same accumulated intelligence. It supports recurring marketing strategy and execution without rebuilding context for every task.
Client histories, treatments, appointments, and follow-ups were scattered across notes, messages, and memory. Important next actions, client opportunities, and rebooking windows were being lost.
I built a system that organizes the complete client operation, maintains the follow-up cadence, and prepares a report before every session. It recovered hours of nightly administrative work and stopped clients and upsell opportunities from falling through the cracks.
“Now it practically runs itself and gives me a report before each session, telling me what needs attention and what is coming next.”
Yes. I work with Claude, ChatGPT, and other LLMs. We can use the platform your team prefers or choose the one best suited to the work.
No. Prompts are only one layer. I build the project instructions, reusable skills, source hierarchy, business rules and exceptions, quality standards, reference knowledge, correction mechanisms, and automations behind the work.
Your team does not need to become good at prompting. They can make a rough or incomplete request, and the system will either execute it according to the rules already built into it or ask for the missing information. Recurring tasks can also run automatically, without a manual prompt.
Simple, focused implementations or single workflows start at $3,000. Once I understand the outcome you want to achieve and what the system needs to do, we define the scope together. I then provide a fixed project price before any work begins.
We start with a conversation about the outcome you want and how the work is currently done. I assess where AI can create the most value and recommend what should be built. If we agree it makes sense, I outline the proposed system and next steps.
You don't need to have it figured out. People reach out to turn expertise into an AI system, to find where AI could create real value, or to fix an AI setup that isn't delivering. The call is for exactly that.
Bring the problem in whatever shape it's in, and we'll work out together whether there's a practical system worth building.
Each page explains one failure pattern and what has to be structured for AI to handle it better.