Integrate generative AI without dependency.
From scoping to deployment: RAG, model selection, security, costs, response quality, and reversibility.
Our way of working.
Independence
Multi-LLM by default. Llama and Mistral in open source, GPT, Claude, and Gemini in proprietary, depending on the task.
Pragmatism
We don’t launch a POC if the data isn’t ready. We don’t push to production if the evaluation doesn’t validate. No reckless forward momentum.
Continuity
AI must integrate with your tools, security rules, and operational practices. Not exist alongside them.
The AI topics we scope and deliver.
Six precise domains where we have deliverables, metrics, and production feedback.
Each use case dictates its choices: model, data, hosting, safeguards, cost.
Internal business assistants
An assistant limited to a clear scope, with sourced answers, refusal rules, and human validation when needed.
Documentary RAG
Connect an assistant to your internal documents to retrieve, summarize, and cite useful information.
Back-office agents
Classify, pre-fill, check, or summarize files, with logs, business rules, and human validation.
Open-source fine-tuning
When prompts or RAG aren’t enough, we specialize a model using validated examples.
Custom MCP servers
Enable an assistant to query your APIs and internal tools—with clear permissions, logs, and limits.
Evaluation & benchmarking
Compare multiple models on your own use cases: response quality, cost, latency, hallucinations, maintainability.
The building blocks we integrate.
None of these are default choices. The stack is decided during scoping, based on your actual context.
Models
Frameworks
Vector stores
Orchestration
Sistr — AI Assistant.
Document search on anonymized patient files. Sovereign hosting on Scaleway, Llama 3.1 70B, RAG with source citations.
Read the full caseA five-step project
A dedicated team from scoping to delivery with a dedicated project manager
Use cases & corpus.
Business feasibility, data quality, sovereignty constraints. Output: POC go/no-go decision.
Argued benchmark.
2-3 models tested on your real corpus. Open source vs proprietary decided on data, not preferences.
The core use case.
2 to 6 weeks. One use case, done right. We validate sources, rights, and assistant limitations.
Quality metrics.
Accuracy rate, cited sources, hallucinations, cost per query, user feedback.
Production deployment.
If criteria are met: deployment, monitoring, documentation, and handover to teams.
Security, compliance, reversibility.
These topics aren’t addressed at the end. They frame decisions from the start: data, access, hosting, logs, monitoring, and exit clauses.
Sovereign hosting
Scaleway, OVH, self-hosted. Data and inference in the EU when required.
Open-source models
Llama, Mistral, Phi. Self-hosted via vLLM or Ollama. No data leakage to third-party clouds.
GDPR & compliance
Privacy by design. HDS-compatible for healthcare. DPIA support if needed.
Reversibility
Code delivered, documented, tested. No proprietary platforms. Everything is retrievable.
Frequently asked questions
The most common questions during scoping. If yours isn’t here, reach out!
75011 Paris