I help teams actually build with AI, designing agents that do real work, automating the busywork, and making the deep concepts click. Everything below is something you can poke at right now.
Agents that plan, use tools, and act on your systems, not just chat. Here's one running a real support-triage task, step by step:
Plan, act, observe loops, tool use, and guardrails scoped to your actual data and systems.
Wiring agents into your APIs, databases, and internal tools so they can actually get work done.
Evals, fallbacks, and human-in-the-loop checkpoints so agents stay trustworthy in production.
The boring, repetitive work, scraping, monitoring, summarizing, reporting, handled on autopilot. Here's one I built and still run today:
A fully autonomous data pipeline: every day it scrapes the price of empanadas at an Argentine supermarket, tracks it against the USD/ARS exchange rate, and updates a live chart. No servers to babysit, no manual steps, it just runs. Click to inspect the real thing.
If a person does it on a schedule and it follows rules, it can usually be automated, reporting, data entry, monitoring, follow-ups.
I make deep AI concepts genuinely click, for engineers, teams, and leaders. Here's a taste: an interactive explainer of how LLMs are actually trained, from scratch.
How large language models are trained, explained with a “100-book universe” analogy, tokens, attention, training, alignment, reasoning models, and more. Written to be succinct but genuinely deep, in plain language.
Four ways I engage, each ending with working software your team owns.
I work directly with founders and small teams to design agentic workflows that take real work off the team, then build and hand them over. No engagement layers, no slide decks you have to translate into software.
Not a platform, and not broad AI consulting. Specific agents that do specific internal jobs, wired into the systems your team already works in, and handed over so your team can run them.
The point of a prototype is to kill or confirm an idea before it consumes a roadmap. I build the smallest working thing that answers the real question, and I tell you what it actually proved.
If you are about to hire an agency to build an AI MVP, it is worth knowing what the alternatives actually trade off. Here is an honest comparison, including the cases where an agency is the right answer.
Free write-ups for the questions people ask me most often.
A decision guide for business process automation: when Zapier, Make, or n8n is enough, and when a custom AI agent or workflow is actually warranted. Includes a comparison table and migration advice.
How to choose between a LangChain or LangGraph consultant and a no-code automation expert for building AI workflows, with practical examples of when code-based agents beat Zapier, Make, or n8n.
Whether it's an agent, an automation, or getting your people fluent in AI, tell me what you're trying to do and I'll tell you if I can help.
Get in touch →Got an idea, question, or just want to say hi?