AI / LLM Engineer senior, applications RAG et agents en production (Python) - Freelance (H/F)

France | Oct. 19, 2026

Report as Closed

Company: Collective.work

Country: France

Type: Hybrid remote

Employment: Full-time

Description: Type of contract: Freelance
TJM: according to profile
Location: West Paris (78), hybrid, 3 days on site and 2 days of teleworking per week
Start: as soon as possible
Duration: 3 months renewable, full time
Client: aeronautics and defense industrial player, name communicated under NDA before the client interview
Prerequisites: French nationality (regulatory constraint of the sector defense)

Context

Large industrial group in the aeronautics and defense sector, which structures its AI team into product squads according to its group AI operational model. Each squad is multidisciplinary and autonomous, responsible for its product from end to end: Product Owner, business contacts, architecture, cybersecurity, application development.

The approach is that of the Minimum Lovable Product: starting from a specific user problem, delivering a first truly useful version, improving it based on use. Quality, safety, traceability and adoption are part of the job from the start. Success is measured by the value obtained by users and the reuse by other teams.

This is not a management mission. The technical referent remains fully in the production: he codes, builds and delivers robust and actually used AI applications, and helps his colleagues progress along the way.

Missions

*

   Design and develop the AI components of the products: document assistants, analysis of technical documents, search and reuse of knowledge, integration of models

*

   Set up agents and orchestrate processing when necessary relevant

*

   Transform a business need into a truly usable application, beyond the demonstration

*

   Build and maintain the AI product evaluation system

*

   Document your technical choices, share your knowledge, support the less experienced

Profile sought

*

   5 to 10 years of experience in software engineering and/or AI, including concrete experience of LLM applications. Neither seniority requirement for LLMs alone, nor upper limit

*

   Decisive criterion: having personally built applications actually used and contributed to their production launch

*

   Software development: Python, API, testing, application integration, containers, CI/CD. The person develops and integrates its components themselves

*

   LLM and RAG: embeddings, vector bases, document cutting, research, reranking, prompt management, orchestration. LangChain, LlamaIndex or equivalent are examples, not exclusive criteria

*

   Evaluation: quality of documentary research, accuracy of answers, citation of sources, hallucinations, robustness, non-regression tests. Knowing how to demonstrate that a modification improves the product

*

   Production culture: integration and versioning of models, performance, inference costs, safeguards, data security

*

   Open weight models: experience sought. Fine-tuning is a controlled option when it is justified, not a systematic response

*

   A career in research is suitable if it is accompanied by demonstrated capacity for achievement

Apply here:

Web: Apply here

Emails: