Lyon has the AI research. Companies are still waiting.

Lyon has an AI research density that few French cities can match. LIRIS lab, Naver Labs (formerly Xerox Research Centre Europe), INRIA, the Mila lab at INSA. Papers get published. Patents pile up. But inside Lyon's mid-sized companies, AI is still a boardroom topic, not a daily tool.

In pharma, the potential is enormous. Sanofi Pasteur in Marcy-l'Étoile generates massive volumes of clinical trial data. BioMérieux processes millions of diagnostic results every year. Pharmacovigilance, drug discovery, clinical data analysis: these are use cases where LLMs already make a real difference. But getting from a validated R&D proof-of-concept to a deployed production tool is a different skill set entirely.

Lyon's industrial base has the same gap. Renault Trucks in Vénissieux could cut machine downtime with predictive maintenance. The Vallée de la Chimie, stretching from Feyzin to Pierre-Bénite, could optimize chemical processes with AI on sensor data. These companies have the data, the engineers, the budget. What they lack is someone who speaks both the business language and the model language.

Then there's the fabric of SMEs and mid-cap companies (ETIs) that makes Lyon what it is. Too small to interest the Big4 for an AI project. Too large to ignore the topic. French Tech Lyon is full of startups that want to integrate AI into their product but get stuck on implementation. The problem isn't the technology. It's the execution.

What I do for Lyon businesses

AI audit: finding the use cases that actually pay off

Working in pharma and want to speed up pharmacovigilance data analysis? Running a plant in the Vallée de la Chimie and looking to optimize processes with sensor data? I look at your data flows, your tools, your regulatory constraints. I identify where an LLM or a predictive model delivers measurable gains. And when a pivot table does the job, I tell you that too.

LLM and API integration: from R&D to production

A RAG system that lets your researchers query thousands of clinical publications in natural language. An assistant that pre-fills pharmacovigilance reports from declared cases. Automated workflows between your LIMS and your reporting tools. Lyon has drawers full of AI proof-of-concepts. My job is to pull them out of the lab and plug them into your business processes, with the security and compliance constraints that come with it.

AI training: making your teams operational

Your data scientists at BioMérieux or your engineers at Renault Trucks don't need a lecture on neural networks. They need to know how to use language model APIs, understand confidentiality limits with patient or industrial data, and prompt effectively in their specific business context. I train on your real use cases, with your data, in your environment.

How it works

1. Initial call (free, 30 min)

We talk about your situation. Your tools, your data, your constraints (regulatory, security, budget). No slides, no pitch. If AI isn't the right answer for your case, I tell you. Nobody's time gets wasted.

2. AI maturity audit

I map your processes and data flows. I identify high-ROI use cases, factoring in your tech stack and sector-specific constraints. You get a clear deliverable: what to do, in what order, with an estimated budget.

3. POC then deployment

We start with one concrete case. A working prototype tested by your teams under real conditions. If it works (and only if it works), we move to production deployment. No six-month development tunnel before seeing a result.

4. Training and autonomy

I train the end users on the tool we built together. Concrete stuff: how to use it, how to write good prompts, how to spot when the AI gets it wrong. When I leave, your team runs it.

Tech stack

OpenAI APIClaude APILangChainRAGPythonn8nGoogle Apps Script

Frequently asked questions

Are you based in Lyon?

No, I work remotely from Tbilisi, Georgia. I grew up in Haute-Savoie, two hours from Lyon, and I know the regional business landscape well. In practice, audits happen over video calls and on your data, development is delivered continuously, and I'm on a European timezone (GMT+4, two hours ahead of France). If an on-site workshop in Lyon is needed, I travel.

How much does an AI project cost?

Depends on the scope. An AI maturity audit for a Lyon-based mid-cap company takes a few days. A full LLM integration with training takes a few weeks. The initial 30-minute call is free and serves to frame the need. I bill by day or by project, with no sales layer or overhead fees. And if the honest answer is 'this isn't worth the investment right now,' you'll hear it.

What about data confidentiality, especially in pharma or healthcare?

That's the first question to address, not the last. I work with private APIs (no data sent to public models), sandboxed environments, and architectures that respect your GDPR and sector-specific constraints. For pharma or health data, we define the confidentiality perimeter before writing a single line of code. No patient or proprietary data passes through uncontrolled services.

My company is an SME. Aren't AI projects only for large corporations?

It's the opposite. Large Lyon corporations have internal data teams and R&D budgets. SMEs and mid-caps have concrete use cases but lack the resources to build an AI team. That's exactly where I come in. One external consultant, a tight scope, a deliverable in a few weeks. You don't need to hire three data scientists to automate your quality reports or connect a chatbot to your technical documentation.