The page most people leave off their website. Most sites only show what worked. This is the rest: two startups that stopped, a mistake I made in my first US meetings, and a visa I treated as paperwork. Each one cost me something. Each one taught me something I now bring to the founders I work with.
Context. A mental-performance platform for tennis players, built with a sports psychologist and a third co-founder. The idea was to combine structured mental training content with a digital platform for coaches and athletes.
What I did. Business structuring, market study, fundraising strategy, roadmap, KPIs, pitch deck, CRM. I built the entire business and go-to-market layer while the other co-founders focused on content and product.
What went wrong. The project stalled on a human problem: intellectual property and control. When it came time to formalize who owned what, the conversation exposed fundamental disagreements about contribution, control, and trust. The same pattern as King Tut, below: three co-founders, same stage, same kind of conflict.
What I learned. The hardest part of a startup isn't the product, it's the people. I now read a founding team before the tech. A contract discussion isn't a formality; it's where the real alignment (or lack of it) becomes visible. I also learned that managing IP in a multi-contributor project requires clarity from day one, not after the work is done.
Context. An anti-counterfeiting technology (Chromostamp) born in a PSL/ENS laboratory. The technology uses a unique physical marker that can be verified with a smartphone. Markets explored: luxury (LVMH, Hermès), mining (gold traceability in DRC), and industrial packaging.
What I did. Built a direct relationship with Bpifrance and brought what I knew about deeptech financing. Built the pitch decks and shared dealflow with investor networks. Presented at the DeepTech Forum (200 attendees) and at EDHEC. Helped structure the go-to-market and advised on fundraising strategy. The project went from minimal communication to 34 pitch decks and a 28-month operational plan.
What went wrong. Internal conflicts within the research team. The lead researcher ultimately wanted to continue doing research rather than build a company. Lab politics, misaligned timelines between science and business, and a fundamental tension between academic caution and entrepreneurial speed. The business co-founder eventually departed, citing a lack of trust and unrealistic expectations.
What I learned. Working with researchers inside a deeptech venture is a specific discipline. The constraints of a laboratory (politics, timelines, risk aversion) are real and must be anticipated, not dismissed. I also learned that the gap between a promising technology and a viable business isn't bridged by pitch decks alone. It requires genuine alignment between every person at the table. This experience directly shaped how I evaluate founding teams in the advisory.
In my first US meetings, I spoke the language but missed the unwritten rules: the subtext, the pace, the polite no. I mistook courtesy for interest. Meanwhile I pitched the French way: grants won, labels and certifications, de-risked milestones.
In front of American investors it read as caution, not conviction. They were asking "why could this get huge?" and I was answering "here is what we have proven." Same facts, opposite framing.
What I learned. The gap isn't linguistic, it's cultural. Translating conviction across the two systems is now the core of what I do, and the first thing I coach.
Want to learn how to talk to US investors? Book a meeting
Context. I finished my master's at Menlo College in May 2026. In April I had filed for OPT, the authorization that lets international graduates work in the US after their degree. In late May I left the country for the summer, with a return flight booked for early October.
What went wrong. By the end of September, my OPT application had been pending for more than five months. That's longer than most published estimates. Over the summer, my passport, with my valid student visa in it, was lost while I was traveling. I got a new passport and applied for a new visa at the US Embassy in Paris. At the end of September, it was refused under section 214(b). That provision is used when the officer isn't convinced of your ties to your home country.
Each piece depended on the next one. No passport, no visa. No visa, no entry. No approved OPT, no right to work once there. One lost document broke the whole chain, and the return flight became unusable. Flights, fees and everything I had to reorganize have already cost me thousands of euros. As I write this in October 2026, it's still not resolved. It's not comfortable to publish. But leaving it out would make this page less honest than it claims to be.
What I learned. I treated immigration as administration to handle once the plan was set. It was the plan. I had planned clients, travel and dates around approvals I didn't control: two agencies and one consular interview. And I left no buffer. A 214(b) refusal isn't a ban: it's a judgment on the evidence presented, and that evidence can be prepared. I now tell founders planning a US move the same thing. Put immigration status on the critical path from day one. Plan in months, not weeks. Know in advance what you would do if a key document disappeared abroad. And talk to an immigration lawyer before committing to anything you can't undo.
The visa routes founders actually use, and what each one allows, are in Decision tools.