Serumon

The autonomous operations layer for cosmetics and personal-care manufacturing.

  • Seed
  • Industrial AI

Overview

Serumon runs the cosmetics plant. Its agents control raw-material dosing and order of addition, hold temperature, shear, vacuum, and homogenisation to build target viscosity and particle size, read shade spectrophotometrically against the standard, govern fill weight across jars, tubes, bottles, and compacts, and drive the robots that cap, decorate, and case-pack. A make-and-fill digital twin simulates the batch before a single kilogram of pigment is committed.

The category has been solved in pieces for thirty years — a vessel controller here, a homogeniser there, a spectrophotometer on a bench, an inspection camera at the end of the line, an MES sitting above all of it recording what already went wrong. Nobody had closed the loop from perception to vessel-and-line control. Serumon does, and every batch it runs makes the next one more accurate.

Léa and Matteo applied in early 2025, three weeks after their first pilot line went autonomous at a contract manufacturer outside Lyon. The plant had cut shade holds to near zero and clawed back overfill giveaway that had been quietly written off for years. They were approved in two days and onto the Industrial & Physical AI track the same week.

What Stackpass unlocked

  • 01

    Compute for a vision-heavy product

    High-speed shade, fill-level, and defect vision is expensive to train and expensive to serve. The maximum cloud and GPU allocation covered both through the first two plants.

  • 02

    Simulation and edge tooling

    Perks across simulation, robotics, and edge-inference platforms let a nine-person team run a digital twin that would normally need a much larger budget.

  • 03

    Regulated-manufacturing mentors

    Sessions with operators who had shipped GMP-grade systems shaped the audit and validation posture before the first enterprise buyer asked for it.

  • 04

    A round built on introductions

    Curated introductions to funds who invest in physical AI, plus round-preparation support, turned a first pilot into a seed round in one quarter.

Founders

  • Léa Vasseur

    Co-founder & Chief Executive Officer

    Formulation chemist by training. Spent nine years leading compounding and shade development for a European contract manufacturer, where she signed off on the batch records that Serumon now writes automatically.

  • Matteo Brunelli

    Co-founder & Chief Technology Officer

    Robotics and controls engineer. Built vision-guided filling and packaging cells for high-speed personal-care lines before turning the same edge stack toward autonomous compounding and rheology control.

“Every other investor asked us for a second pilot. Stackpass asked to stand on the line at four in the morning and watch the batch run. They understood the product before they understood the model.”

Léa Vasseur — Co-founder & CEO, Serumon

Give your startup an unfair advantage.

One application. Credits, tools, mentors, capital intros, and a global founder community — free for qualifying startups, and we never take equity.