<IA/>

AI / ML ENGINEER

Open to opportunities

Artificial intelligence, engineered to hold up in production.

I design intelligent pipelines, multi-agent architectures and predictive models shipped to production, translating real business needs into concrete AI solutions.

00
Projects
00
Positions
00
Certifications
LangGraph · agentqdrant://prod
LIVE

REQUESTS / S

42

LATENCY P95

214ms

ACTIVE AGENTS

5

DOCS INDEXED

8.4k

graph.reasoning

ARCH · INGESTION-AGENTS-MEMORY-RESPONSE

iteration 419

00:00:00INFspawn agent · researcher-01

00:00:00OKqdrant://prod connected · 2.4M vectors

00:00:00REQPOST /v1/embed · 1,536 tokens

00:00:00INFretrieval · qdrant · score=0.91

00:00:00OKreranked 18 docs · top-k=6

00:00:00REQPOST /v1/query · context=8k

00:00:00OKgraph-constrained reasoning · 3 hops

00:00:00INFagent.route · classifier → extractor

qdrant://prodbuild v1.4.2fastapi:8000 ✓00:00

EXPERTISE

Expertise & technologies

Languages & frameworks

/ 06
PythonJavaScriptDjangoDjango REST FrameworkFastAPINext.js

Generative AI & agents

/ 06
LangGraphLangChainLangFuseCrewAIRAGSystèmes multi-agents

ML & data science

/ 05
NumPyPandasscikit-learnHugging FaceTensorFlow

Data & infra

/ 04
PostgreSQLRedisMongoDBQdrant

MethodsAgile / Scrum · Analyse des besoins métier

// STACK_ACTIVE

live

Stack in motion

AI
tools_in_motion = 12updated continuously

PROJECTS

Selected projects

View all projects
01
Conversational AI & RAG · 202601

ARIS

Constrained

REASONING

Semantic

SEARCH

Traceable

ANSWER

A chatbot that answers with the institute's own publications, researchers and patents — never at random.

  • Graph-constrained reasoning to avoid hallucinations
  • Semantic search over the institute's research ontology
  • Answers traceable back to their sources
PythonLangGraphLangChainQdrant
02
ERP & workflow engine · 202602

E-OPS

Modular

ARCHITECTURE

AML/CFT

COMPLIANCE

Single

HISTORY

A modular ERP that digitizes company procedures, from AML/CFT compliance to procurement.

  • Orchestration engine shared across every business module
  • Opt-in modules: HR, finance, procurement, document management
  • One auditable history, built for AML/CFT compliance
PythonDjangoDjango REST FrameworkPostgreSQL

METHOD

How I build

Turning an uncertain business need into an AI system teams can deploy, measure and evolve.

01

Start from the business need

02

Design with the right, reliable tools

03

Ship to production