Federico De Ponte · Founder, Floom · Founders Inc SF
I build the agent infrastructure that turns teams AI-native.
mex, 2024
sf, 2026
berlin, 2025
Lovable generates beautiful code. Having agents do your actual work is the real magic: research, pricing, QA, outreach, and ops running every day with logs, approvals, and owner-grade reliability.
Proof of work
$600K
ARR at Scaile after rebuilding the operating model from 10 FTE to 2.5 while tripling revenue
20K+
Open-source installs across AI workflow tools shipped from founder pain, not demo-day theater
20
Agents running daily on my own stack for research, QA, product, content, and internal ops
Ex-BCG
Strategy at BCG and digital pricing at Kühne+Nagel before building companies
What I do
I help post-seed teams move from AI experiments to agent-run operations. The work is not a chatbot layer. It is production infrastructure where agents gather context, call tools, draft decisions, ask for approval, and leave an audit trail.
Best fit: workflows that still live across spreadsheets, Slack, docs, browser tabs, and one person's memory. I map the work, remove the brittle steps, and ship the interface your team uses: a live URL, API endpoint, queue, logs, and handoffs.
Current build
Floom is the product version of that work: infrastructure that turns Python functions into live web apps, REST APIs, and MCP tools for teams and agents. See it here
The promise is simple: write the function, get the URL. A script becomes a working tool in under a minute, with a surface humans can use and agents can call. That is the deployment layer every AI-native team needs.
Selected work
A few systems behind the point: AI visibility for enterprise GTM, startup job matching, long-form research agents, and the deployment layer for internal AI tools.
Signals
IFB Hamburg InnoFinTech
NCA backed
BCG alumBuild with me
If your team repeats the same messy workflow every week, send it over. I'll tell you what can become agent-run, what needs human judgment, and what I would ship first.