Fractional Deputy CEO · Deep-tech · Biotech

You built the technology.
I'll build the company around it.

Finance, HR, Legal, Admin, IT, and Regulatory — owned and delivered. Not advised.

Book a discovery call See what I cover →
10+
years in deep-tech ops
3–36
month mandates
A→C
series covered
100%
operational ownership
Why it matters

Operations don't run themselves. But they shouldn't run you either.

Every month a founder spends on payroll, compliance, vendor negotiations, and IT firefighting is a month not spent on the product, the science, or the market. The cost is invisible — until it isn't.

Finance consumes founders

Cash flow gaps, CAC accounting, statutory filings, grant reporting — individually manageable, collectively exhausting. One missed BPI reporting deadline costs more than six months of a fractional CFO.

Regulatory and legal drift

IP not registered at the right moment. A KBIS anomaly spotted in a data room. A contractor misclassified. In deep-tech, the operational fabric of the company is also its valuation narrative.

The headcount dilemma

A full-time Deputy CEO is expensive and hard to reverse. A consultant gives advice you have to implement yourself. A fractional operator takes ownership — and hands it back when the company is ready.

Scope of work

Everything between the product and the cap table

I take full ownership of the internal functions that don't generate revenue but determine survival. Reporting to the CEO, accountable for results.

📊

Finance & Cash Management

Cash flow modelling and forecasting. P&L supervision. Purchasing optimisation. CapEx planning. Statutory accounts coordination. Bank relationships.

🏛️

Non-dilutive Funding

BPI France, ANR, EIC Accelerator, Horizon Europe, CIR/CII — sourcing, structuring, writing, and reporting. From first application to audit defence.

👥

HR & People Operations

Hiring process design. Employment contracts under French labour law. Onboarding. Performance frameworks. Payroll coordination. Labour compliance and social audit prep.

⚖️

Legal, IP & Regulatory

Contract lifecycle management. IP filing strategy and portfolio oversight. Regulatory pathways (CE marking, FDA pre-submission). Fundraising legal coordination.

🗂️

Administration & Office

Back-office structuring. Vendor and supplier management. Facilities and leases. Insurance. Corporate housekeeping (KBIS, statutes, AGM).

🖥️

Technology & IT

Infrastructure and sysadmin. Security baselines and access management. AI workflow automation. Data governance and GDPR compliance.

⚙️

ERP & Tooling

Odoo implementation and customisation. CRM, PLM, QMS integration. Tool rationalisation and migration planning.

🤝

Team & Contractor Management

Org design and hiring roadmap. OKR framework implementation. Subcontractor supervision and audit. Supplier KPIs. Management of management.

⚡ Run & Build — simultaneously. Operational continuity (run) and transformation projects (build) managed in parallel. Because a company at Series B cannot stop while it reorganises.
Engagement model

Three steps to embedded leadership

01

Discovery

A 30-minute call to map your operational gaps, priorities, and timeline. No pitch, no deck — a structured diagnostic conversation.

02

Scoped proposal

A written proposal specifying scope, KPIs, reporting structure, and duration. Fixed mandate or rolling monthly — your choice.

03

Embedded mandate

I join the team, take ownership of the agreed functions, and operate with the authority of a C-suite member. Weekly reporting to the CEO.

Typical mandate length: 3 to 18 months. I carry a maximum of two mandates simultaneously to ensure full availability.

Editorial line

Evidence-based management

Management is full of practices selected on the basis of narrative, intuition, and peer pressure. I apply to operational decisions the same standard I would apply to a clinical study: what does the replicated evidence actually say?

This is not a rejection of judgement — it is a demand that judgement be calibrated. Assumptions stated. Priors updated. Effect sizes checked against operational relevance, not just statistical significance.

Effect sizes, not anecdotes

Business biographies are not evidence. Case studies are hypotheses at best. Interventions are evaluated on replicated effect sizes across heterogeneous samples.

No unfalsifiable frameworks

Personality assessments with no predictive validity. Transformation programmes with no measurable outcome. Culture decks accountable to no one. These do not have a place here.

Bayesian updating

Prior beliefs are stated explicitly and revised proportionally to the strength of incoming evidence. The first study is a hypothesis. A tenth independent replication is closer to policy.

Clinical significance over p-values

"Statistically significant" does not mean operationally meaningful. A 0.1% absolute risk increase that "doubles your risk" is not worth reorganising your company for.

Replication before adoption

Initial effect sizes systematically decline with subsequent replications (the decline effect). We do not implement practices because a TED talk was persuasive.

Publication bias is the baseline

Positive results are published. Negative results are filed. In management research as in medicine, the visible literature overstates effects by construction. Adjust accordingly.

Read the full argument on the blog →
Myself
Lionel Arnaud
Lionel Arnaud
Operations leader · Biotech & Deep-tech
View LinkedIn profile
BPI France EIC Accelerator Agoranov Odoo French labour law IP strategy Clinical trial design AI automation Paris

A decade of operational leadership in biotech and deep-tech — including Biomemory (DNA data storage), where he built the company's operational infrastructure from early stage. Expert advisor for Agoranov (Paris's leading deeptech incubator) and the European Innovation Council EIC Accelerator programme.

Trained in clinical trial design (Johns Hopkins University) and computer science (Harvard CS50X), with a postgraduate degree from Université Paris Dauphine. The same epistemological rigour he applies in scientific due diligence — distinguishing prior probability, effect size, and replication from statistical noise — is what he brings to operational mandates.

Bilingual French–English. Based in Paris. Available for remote and on-site mandates across France and Europe.

Next step

Book a discovery call

Thirty minutes. No deck. A structured conversation about where your operations are today and what they need to look like in six months.

Schedule a 30-min call
or write to
contact@dinnizer.com I respond within one business day. Calls available in French and English.
Questions

Common questions

Coming soon

OpenClaw — Internal Management OS

A full company operating stack built in TUI/CLI. Data owned, portable, and available to train private AI models. Because your company's operational data should never live in a vendor's cloud.