Applied AI
AI in your workflows, down to the hardware.
I connect AI to your tools, documents and processes, and down to your sensors, machines and test benches. For freelancers, small businesses and SMEs.
01 / What I can do for you
Real use cases, not demos.
- 01
Document processing
Extract, classify and summarize invoices, contracts, reports or specifications, then feed the results into your systems.
- 02
Internal assistants
An assistant that answers from your documentation, procedures and tickets, and cites its sources.
- 03
Workflow automation
Chain the steps between email, ERP, CRM, spreadsheets and databases; AI handles the unstructured parts.
- 04
Agents & MCP tools
Agents that act on your tools through controlled interfaces (MCP servers, APIs), with permissions and logs.
- 05
Field data
Sensors, machines and test benches connected to your tools: AI turns their measurements into reports, alerts and decisions.
- 06
Technical teams
Reading datasheets, bills of materials (BOM), test log analysis, documentation: what I use myself every day.
02 / The right model for each task
No lock-in to a single vendor.
I pick the model by task, cost and confidentiality: frontier models from OpenAI or Anthropic for the hardest reasoning; Jev, from TypeSafe AI, for typed decisions at the heart of automations; open-weight models such as DeepSeek, Qwen or Kimi, which can run on your own servers when data must stay inside the company.
- OpenAI
- Anthropic
- DeepSeek
- Jev
- Kimi
- Qwen
Logos are trademarks of their owners; they show the models I work with, without any affiliation.
03 / How a project runs
- 01
Audit
I map your repetitive tasks and measure the time they cost.
- 02
Prototype
A first result on your real data, within days.
- 03
Integration
Into your tools, with tests, guardrails and human validation.
- 04
Measurement
Time saved, error rate, cost per task: numbers, not impressions.