Capabilities

What I do, in concrete terms.

Five areas of expertise to go from idea to a system in production. You come in through what you know how to do, or through the problem you want to solve.

01

LLM & RAG

Your documents become a knowledge base you can query in plain language, with sourced answers.

The right answer, sourced, in seconds.

2 s

for a sourced answer from your documents

100 %

of answers cite their sources

0

made-up answer tolerated in production

What I build

01

Document RAG

Contracts, procedures, records: your documents become queryable in plain language, with source citations.

RAG · Sourced answers · Semantic search

02

Search & extraction

Finding, extracting and structuring information buried in your files, emails and databases, at scale.

Extraction · Structuring · pgvector · Qdrant

03

Model selection and integration

The right model for each task (Claude, GPT, open-source), cleanly plugged into your systems.

Claude · OpenAI · Open-source · MCP

The stack

  • Claude
  • OpenAI
  • RAG
  • pgvector
  • Qdrant
  • Embeddings
  • Reranking
  • Python

02

Agents & orchestration

Agents that carry out real tasks in your tools, coordinated when one isn't enough.

An agent that acts, not a chatbot that talks.

24/7

the agent never sleeps, never forgets

multi-agent

each with its own task, orchestrated

MCP

your tools cleanly connected

What I build

01

Tooled business agents

An agent that answers, decides and executes actions in your tools, and escalates to a human when needed.

Tool actions · Human escalation · Support

02

Multi-agent orchestration

Several specialized agents that coordinate on complex tasks, with an orchestrator and guardrails.

Multi-agent · Coordination · Guardrails

03

Connecting to your tools (MCP)

Your CRM, emails, databases and APIs connected to the agent through clean, maintainable integrations, including MCP.

MCP · Integrations · API

The stack

  • MCP
  • LangGraph
  • Claude
  • Tool use
  • n8n
  • Python
  • TypeScript
  • Webhooks

03

Cloud & on-prem

Your AI systems deployed wherever your constraints require, down to your own infrastructure, without your data ever leaving.

Your data stays with you.

on-prem

private models inside your infrastructure

EU

sovereign hosting possible

0

data leaking outside your perimeter

What I build

01

Cloud deployment

Scalable, monitored production deployment on your cloud (AWS, GCP, Azure, Vercel), with CI/CD and controlled costs.

AWS · GCP · Vercel · CI/CD

02

On-prem & private models

Open-source models deployed inside your infrastructure: your data never leaves your perimeter.

On-prem · Open-source · Sovereignty · Security

03

Industrialization (MLOps)

Versioning, monitoring, zero-downtime updates: an AI system you can run with peace of mind over time.

MLOps · Monitoring · Zero-downtime

The stack

  • Docker
  • AWS
  • GCP
  • Azure
  • Ollama
  • vLLM
  • Terraform
  • Vercel

04

Audit & optimization

I find where AI truly creates value, and cut the bill on what's already running.

AI at the right price, where it pays off.

−30 to −70 %

on your AI bill, at the same quality

ROI

every use case quantified

48 h

for a first audit readout

What I build

01

AI audit & strategy

Your processes put under the microscope to find the use cases with real ROI, and honestly rule out the ones that aren't worth it.

Audit · Quantified use cases · Roadmap ROI

02

Cost optimization

Model selection, caching, prompt architecture, batching: the LLM bill goes down without losing quality.

−30 to −70 % · Caching · Batching · Model selection

03

Feasibility scoping

Before you invest: what's feasible, at what cost, with what risks, on your real data.

Feasibility · Targeted POC · Risk analysis

The stack

  • Token audit
  • Prompt engineering
  • Caching
  • Batching
  • Observability
  • Evaluation

05

Reliability in production

Evals, guardrails and observability for AI systems that don't drift and don't break silently.

AI that doesn't go off the rails.

evals

every version measured before production

guardrails

drift blocked, not endured

observability

you see what the agent is doing

What I build

01

Evaluations (evals)

Test suites that measure the real quality of your answers, on every version, before it ships to production.

Evals · Regression testing · Quality

02

Guardrails

Filters, output validation, action limits: the agent stays in its lane, even when facing the unexpected.

Guardrails · Validation · Security

03

Observability

Traces, logs, alerts: you see what the system is doing in production and you're warned before it drifts.

Tracing · Monitoring · Alerting

The stack

  • Evals
  • LangSmith
  • Traces
  • Guardrails
  • Observability
  • Logfire
  • Monitoring

Contact

Ready to build something that matters?

Book a call

Reply within 24 h · first conversation free with no strings attached.