CTO

Il y a 3 jours

Paris, Île-de-France Sonalyze Temps plein

CTO Wanted – Sonalyze @ Station F x Ville de DemainTable of Contents


Si vos compétences, votre expérience et vos qualifications correspondent à celles de cet aperçu de poste, n'attendez pas pour postuler.
  • A new kind of "DPE" - for sound
  • Where we stand - this is not a napkin-stage project
  • Who I'm looking for
  • What we'll build together
  • What you get
  • Interested?

I'm looking for a CTO co-founder - and not a classic one.

I'm looking for a maker, a hacker, an embedded / edge-AI savant who wants to take a product Europe has never seen from the working prototype phase to industrial reality while using a full agentic workflow easing the workload by *100.

Who we are

Sonalyze was created in March 2025 and is incubated at Station F (Ville de Demain program). In 17 months, with a frugal team, we built a full product prototype: our own measurement device, the ML pipeline and the platform, all working end to end.

We showcased it successfully at VivaTech (with the CCI), at Panathēnea in Athens and at LIDD in Lille. Next stops: the national social-housing congress in Bordeaux (September), RENT Paris (November) and ANQES, the national conference on sound environment quality, in Le Mans (November).

We are backed by the CCI startup program and by the Scaleway and OVHcloud startup programs. Cloud credits, network, visibility: the ecosystem is in place. What's missing is you.

A new kind of "DPE" - for sound

In real estate, we certify energy performance, insulation, structural integrity... But nobody can reliably diagnose the sound performance of a home.

Yet noise is the #1 nuisance reported by residents. Every year, millions of euros are lost in noise disputes, renovation surprises, tenant conflicts and legal procedures. Buyers, landlords, social housing operators, property managers, insurers, cities — they all need trustworthy acoustic diagnostics.

Sonalyze, incubated at Station F (Ville de Demain program), is building exactly that:

the DPS - Sound Performance Diagnostic - for every home and building, a measurement device + embedded AI + an A-to-G rating, like the DPE but for acoustic comfort, built on the vocabulary of the real standards (IEC 61672, ISO 1996-2, ND 2002/49/EC (END: Environmental Noise Directive)) so professionals take it seriously.

Nobody has built the reference for this. Not in France, not in Europe, not anywhere. We intend to be it.

Where we stand - this is not a napkin-stage project

  • Sona One, our own board, is fab-ready. Proprietary 6-layer PCB for embedded AI, on battery, LTE and USB-C and our MEMS microphone. Full CTO-grade design audit passed, EMS quotes in hand, pilot assembly launching now.
  • Our own enclosure, designed for measurement: IP54 target, "class-2-like" acoustic path, calibrator adapter, tool-free removable battery (EU battery regulation 2023/1542-ready). First MJF print run in progress.
  • The DPS methodology exists: an A–G scale with defined thresholds, a standards referential (IEC 61672 / 61260, ISO 1996-2, ISO 20906, EU END) mapped version by version, and a written robustness test campaign.
  • The AI is real: our sound-classification and acoustic-metrics pipeline has already run 24h inferences end-to-end with objectif of 1 week; it is now being ported to our own silicon. Privacy by design — raw audio is never stored.
  • One European patent filed (EPO), international extensions planned.
  • Money is in motion: JEI status in progress, Bpifrance deeptech grants filed, EIC Accelerator in progress, seed round planned for end 2026.
  • The market is answering: exhibitor at the national social-housing congress in September, real-estate innovation events and competitions this autumn, public-procurement referencing underway.
  • Cost target: ~€200 hardware + €200 service - a mass-market instrument, not a lab toy.

And the use cases go far beyond real estate: bailiffs, cities, social landlords, insurers, construction and renovation - and at term a low-cost, high-volume European acoustic-intelligence infrastructure with real world acoustic data.

Who I'm looking for

Not a "manager CTO." Not a delegator. Not a PowerPoint architect.

Someone who:

  • lives in embedded C/C++ and Python, and reads a schematic without flinching;
  • knows edge AI for real: model compression techniques (prunning, quantization), TinyML constraints, compute, memory and power budgets;
  • has opinions about DSP, sampling and filters — or must master acoustics fast;
  • can take a board from bring-up to certified product: drivers, RTOS, self-test, OTA, CE marking (RED / EN 18031);
  • builds fast, tests fast, iterates fast — and finishes;
  • thinks in TRL, not vanity metrics;
  • is in Paris and wants to sit at Station F with me — not on a Google Meet.
  • Rust is a plus
  • L