Freelancer profile translated to English.
Back to original languageAbout Etienne
🧠With 5 years of experience in Data/AI, I have built a professional roadmap marked by diverse environments: large corporations, start-ups, and independent missions.
I have contributed to various projects ranging from rapid prototyping in healthcare to the design and industrialization of critical pipelines in banking.
🚀 Recently certified AWS Machine Learning Specialty, I maintain continuous learning, with particular attention to advances in GenAI. My approach combines technical rigor, product vision fueled by my Agile/Scrum experiences, and the ability to translate complex business problems into concrete solutions.
🤝 Looking to explore a Data/AI challenge? I would be delighted to chat to see how we could move forward together.
🎯Skills:
🔹Development: Python, API, Docker, MLOps, CI/CD (GitHub Actions)
🔹Cloud: Amazon Web Services (AWS), GCP
🔹Databases: PostGreSQL, DynamoDB, Neo4j
🔹Pipeline Orchestration: Airflow, Sagemaker, StepFunctions, Kubernetes/EKS
🔹Machine Learning: AI Search, Fraud Detection, NLP, Computer Vision, ASR
🔹Data Transformation & ETL/ELT: Pandas, AWS Glue, Apache Spark, dbt
🔹DataViz: Power BI, Tableau, Python (Matplotlib, Seaborn)
🔹GenAI/LLM Agents: LangChain, MCP, RAG, vLLM, AWS Bedrock, PromptLayer
🔹Product Development: Scrum/Kanban, KPI/OKR, Prioritization (RICE/MOSCOW)
✍🏻Service Offering:
🔸 Data Readiness Audit/Workshop
🔸 AI Project Scoping, Use Case Definition, AI ROI Workshop
🔸 End-to-end Data Science/AI Project Development (Design / MVP to Industrialization / Production)
🔸 Data/AI Project Management / Run
French
Native or bilingual
English
Native or bilingual
Can work on-site
Paris (up to 30km)
Experience
- Roc EclercMachine Learning EngineerPUBLIC SECTORDecember 2024 - February 2025 (2 months)Paris, FranceContext:During the pre-sales phase for a funeral industry player, the objective was to demonstrate the value of anadvanced AI search barcapable of improving user information gathering and encouraging their entry into the conversionfunnel(click on quote request after a search).
- Challenge: Their competitor had already implemented a search engine, like Algolia/TypeSense.
- Identified **Problems**: High bounce rate, confusing navigation from the landing page, improvable SEO.
- Assumed key user intentions: Funeral preparation and foresight.
Output:Comparative search engine prototype- Benchmark and demo of three approaches: Syntactic engine (keywords), semantic (embeddings), hybrid (BM25 + embeddings, ANN).
- Indexing and ranking via ElasticSearch, allowing for open-ended questions and consistent results.
Search Features- Development: Custom crawler, indexing pipeline, API with FastAPI, monitoring with Kibana, Docker.
- Faceted navigation, autocomplete, debouncing, zero-result fallback, basic weighting ranking.
Product Scoping- Organization in Kanban board, weekly prioritization with the Head of Product and an Account Manager.
- Hypothesis: a relevant search bar → reduced bounce rate and better conversion.
- Next Step: Gather user feedback (e.g., keywords vs. open-ended questions) to "prime the pump" and implement business/user metrics
Outcome- Successful **Proof of Value**: Convincing internal demo of the hybrid engine's superiority, validating the business interest of an advanced AI search.
- Highlighted potential improvement in CTR and retention
- The POV allowed a PO (Wivoo) to capitalize on the work done.
- ThIA Santé MentaleMachine Learning EngineerHEALTH AND WELLNESSOctober 2024 - December 2024 (2 months)Montpellier, FranceContext:Mission within ThIA, an e-health start-up developing a medical monitoring platform and collecting data from Doctolib/AviPsy.
- Responsible for thesizinganddesignfrom scratch of their Data architecture (Data Lake & Warehouse HDS) and an analyticalELTpipeline.
- Data Mart designed to meet the needs offoresightstakeholders: operational monitoring, patient journey analysis, and regulatory reporting.
- Direct collaboration with the CTO: daily meetings, backlog refinement, and sprint planning.
Output:Architecture & Data Pipeline- Construction of an event-driven pipeline (DataFlow, **Airflow**, Pub/Sub) to process heterogeneous sources (CSV, REST API, PostgreSQL).
- Deployment on GCP in a medallion architecture (CloudStorage, BigQuery, dbt).
- Docker, CI/CD GitHub Actions
- Implementation of robust mechanisms: retry, rollback, data and schema versioning. Minimization ofdowntime(SLA > 99.99%).
Governance & Monitoring*Data Catalogand data lineage management to ensure traceability and compliance (HDS/GDPR).- Granular access controls via IAM and Secret Manager, systematic anonymization of patient data.
- Cloud monitoring: DAG tracking, resource consumption, and intelligent **alerting**.
Analytics & Reporting- Standardization of transformations via dbt macros.
- Implementation of thematic analytical Lookerdashboards(practitioners, patients, regulatory).
- Pipeline and infrastructure sized to accommodate advanced use cases: Modeling the risk of insurance contract non-renewal (**churn**)
Outcome:- Industrialization of the pipeline and delivery to production in a regulated environment.
- Foundation laid for analytical and predictive use cases in the health/insurance sector.
- Société GénéraleData Scientist ConsultantBANKING AND INSURANCEFebruary 2022 - June 2024 (2 years and 4 months)Fontenay-sous-Bois, FranceData Scientist / Engineer – Network Flows & Fraud DetectionContext:Service provided within the IT Innovation and Business Technologies IT department, as part of an anti-fraud program for **ransomware attack prevention**.Contribution to a Big Dataagileproject from scoping to **Production**, in continuous interaction with technical architects and functional network experts.**Objective**: Map network flows between applications and users to improve and enrich **business KPIs**, reduce dangerous protocols, and strengthen data governance.Output:ETL pipeline design and management
- Design of acomplete ETL workflow(firewall logs, repositories, metadata) with business stakeholders (Domain-Driven Design approach)
- Parsing, deduplication, source consolidation, massive aggregations (**Spark**), monitoring, and reporting
- **Governance**: Alignment with modeling needs, **lineage**, and audit compliance for each use case
- Deployment of aBusiness Rules Engineintegrating business rules per environment (IaC)
- Technical lead: proposing and arbitrating technical choices
Industrializationand R&D work- Provision ofData Martsfor business stakeholders (RESTful APIs, parametric SQL)
- Supervision of codebase porting (Python → Scala) and CI/CD pipeline
- Supervision of agraph POCfor interactive exploration of flows between applications and partners
- Implementation of Graph Neural Networks (DGL) to detect advanced networks of "mule accounts", not captured by traditional methods
Outcome:- Complete coverage of the bank's network interconnections (~100GB/day)
- Production deployment of an orchestrated analytical product, essential for business teams and with group-wide scope
- Operation: Historical catch-up + Bi-daily scheduled run
Recommendations
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Master's Degree, Open Information Systems Engineering (SIO), SaclayCentraleSupélec2020Mastère Spécialisé, Ingénierie des Systèmes Informatiques Ouverts (SIO), Saclay
- Software EngineeringESILV - Ecole Supérieure d'Ingénieurs Léonard de Vinci2019Software Engineering
Certifications
- AWS Certified Machine Learning – SpecialtyAmazon Web Services2025
- Professional Machine Learning Engineer CertificationGoogle Cloud2024