Freelancer profile translated to English.
Back to original languageAbout Louis
AI OCR pipeline in production (40+ users), multi-company RAG system, multi-tenant SaaS delivered in 2 weeks. I build production-ready AI agents and fullstack applications.
My skills:
- Applied AI integration (local LLMs, RAG, OCR, document automation) with gains of up to 20 min/day/user.
- AI workflow integration (Claude API, n8n, Zapier, Langchain)
- Fullstack React / Next.js / Node.js development (REST API, JWT, scalable architecture, TypeScript).
- Backend Node.js & Django, secure API design and complex business logic management.
- Architecture & deployment Docker, Linux VPS, CI/CD, production environment.
- Web performance optimization, Core Web Vitals, technical SEO (Lighthouse 94/100 perf, 100/100 SEO & accessibility).
- Magento 2 / Hyva development, custom B2B modules, advanced business rules.
- Design of robust, maintainable, production-oriented, and scalable applications.
- Business needs analysis → technical translation → production delivery.
- Agile environment, Git, code review, software quality, refactoring.
My approach: understand the business need before the stack.
I can work on a complete architecture or a critical, high-value module.
French
Native or bilingual
English
Fluent
German
Basic
Spanish
Basic
Can work on-site
Paris (up to 50km), Strasbourg (up to 50km)
Experience
- S2 DEVELOPPEMENTAI Project ManagerTRAVEL AND TOURISMJune 2026 - Today (2 months)Mulhouse, FranceContext: A multi-agency group receiving a flow of incoming requests to qualify. Objective: score each incoming lead, enrich it as much as possible, and route it to the correct agency or the group's BizDev.Resolution & Enrichment:Multi-agent pipeline resolving the legal entity (SIRET/SIREN) from minimal inputs, with separate collector and judge agents (the judge scores confidence in the establishment, funnel logic: stop at max confidence, otherwise broaden tools).Multi-source scraping (legal notices, LinkedIn contacts, executives, headcount) and firmographic enrichment: 3 fiscal years financials, group structure, business qualification, origin by field.Scoring, Qualification & Routing (Product Core):Multi-dimensional scoring (deal/account/contact) with grades and fit × engagement quadrant to prioritize incoming requests.Routing new → qualified → routed → contacted, land-and-expand signal (small entry ticket on a large group → BizDev alert), identity inconsistency flag (headquarters vs. local branch).Scoring consistency audit: aberrant correlations (priority driven by account not deal), scoring on unresolved identities, misaligned grade thresholds.Asynchronous orchestration BullMQ/Redis, multi-tenant dashboard (agency/group RLS), GDPR compliance on scraped PII.Stack: Python, Next.js, TypeScript, PostgreSQL, Prisma, BullMQ, Redis, Supabase, pgvector
- Bellum AIAI EngineerHUMAN RESOURCESJune 2026 - Today (2 months)Paris, FranceContext: B2B account acquisition AI agent for an IT consulting firm. A salesperson enters a company name and obtains a complete, personalized, reliable, and sourced account dossier: who to target, how to approach, what message, when.Resolution & Enrichment (Product Core):Multi-agent pipeline for legal entity resolution (SIRET/SIREN) from minimal inputs (email + partial name + city), robust to edge cases: trade names, holding companies, sole proprietorships, deregistered companies, false positives from website publishers. Confidence judge agent with funnel logic (stop at max confidence, otherwise broaden tools). Multi-source enrichment (INPI/RNE, SIREN, LinkedIn) with origin by field. Asynchronous orchestration BullMQ/Redis: idempotent checkpoints, typed errors, retry/DLQ, contract tests.RAG Engine:Hybrid search pgvector + PostgreSQL full-text merged in RRF, listwise reranking (Claude Haiku). Injection of an IT consulting firm profile in a 4-phase pipeline, with strict anti-prompt injection boundaries.Scoring & Routing:Multi-dimensional scoring (deal/account/contact) with grades and fit × engagement quadrant. Sales pipeline (new → qualified → routed → contacted → won/lost) and land-and-expand signal. Consistency audit: aberrant correlations, scoring on unresolved identities, misaligned thresholds, headquarters vs. branch inconsistencies.Multi-tenant Next.js (RLS isolation), GDPR compliance on scraped PII.Stack: Next.js, TypeScript, PostgreSQL, Prisma, pgvector, Python, BullMQ, Redis, Supabase, RAG, Claude Haiku
- DirigoFullstack DeveloperTRANSPORTATIONMarch 2026 - Today (5 months)Bordeaux, FranceContext: ERP editor specializing in operational management for transport and collection companies. Maintenance of the historical FileMaker ERP used by 500 to 1,000 end-users, and development of the new generation of the product: a multi-tenant SaaS in Next.js.Team: Direct collaboration with the CEO, complete technical autonomy.FileMaker ERP (maintenance and evolutions):Integration of DocuSign for electronic contract signingAutomation of DPAE (Declaration of Hiring) and employee contract generation for clientsFunctional evolutions based on field feedbackSaaS Dirigo (design and development of the target product):SaaS management platform covering routes, weekly schedules, employees, vehicles, leave, and invoicing.Multi-tenant architecture with complete isolation per tenant_id (Prisma extension, tenant resolution by subdomain, stateless JWT)RBAC with 6 roles and 65 granular permissions with hierarchical scope (Center → Team → Lot → Route)Automatic weekly schedule generation algorithm (permanent staff, flexible staff, conflict detection, regeneration on event)Leave approval workflow and complete audit trail on all changesRobust test suite: 869 unit tests, 61 E2E tests, ~76% coverageResults:Historical ERP stabilized without service interruption for the 500 to 1,000 end-usersEnd-to-end automated HR processes for client carriersSaaS deployed in production, positioned as Dirigo's next-generation productStack: Next.js 15, TypeScript, PostgreSQL 16, Prisma 6, NextAuth.js, shadcn/ui, Tailwind CSS, Zod, Vitest, Playwright, Docker, OVH, FileMaker, DocuSign API
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
- BUT Informatique (University Technology Diploma)IUT Robert Schuman Illkirch2025Cette formation de référence m'a permis d'acquérir une expertise pluridisciplinaire en ingénierie logicielle. J'y ai développé une solide capacité d'abstraction et une méthodologie rigoureuse pour concevoir des solutions informatiques performantes et évolutives. Compétences cœur de métier : Développement Fullstack : Maîtrise des cycles de développement, de la conception de la base de données à l'interface utilisateur. Génie Logiciel : Application des bonnes pratiques de code (Clean Code), gestion de versions (Git) et architectures logicielles. Gestion de Projet Agile : Capacité à travailler en équipe et à livrer des incréments logiciels fonctionnels dans des délais impartis. Projets & Livrables : Conception et réalisation d'applications web et mobiles robustes. Modélisation de bases de données complexes et optimisation de requêtes. Valeur ajoutée : Un socle académique technique de haut niveau qui garantit la fiabilité et la qualité du code produit sur chacune de mes missions.
- IAS Level 2 (Insurance Intermediary)Prodemial Buisness School2025Cette certification dote mon profil d'une expertise réglementaire solide pour l'intermédiation en assurance. Elle valide ma capacité à conseiller des solutions de protection et d'investissement tout en respectant les cadres juridiques et déontologiques les plus stricts. Compétences certifiées : Conseil Réglementé : Maîtrise des produits d'assurance vie, de prévoyance et des solutions d'épargne. Analyse de Risques : Capacité à évaluer les besoins de protection des clients pour proposer des couvertures sur mesure. Conformité (Compliance) : Application rigoureuse des normes ACPR et des devoirs de conseil et d'information. Valeur ajoutée : Garantie d'un accompagnement professionnel, éthique et conforme aux exigences du marché financier. Double casquette technique et réglementaire pour sécuriser les projets d'investissement.