Backend RAG & Agentic AI Engineer (M/F)

France | Nov. 12, 2026

Report as Closed

Company: Act Digital France

Country: France

Type: Hybrid remote

Employment: Full-time

Description: act digital is an international consulting and engineering group that supports its clients in their digital transformation projects. Present in 12 countries and with more than 7,000 employees, we put our expertise at the service of their challenges in software development, cybersecurity, data, cloud and AI. Our ambition: to become the trusted technological partner of the most innovative companies, by designing and securing systems that strengthen their performance and resilience. Joining act digital France means joining an agile and committed, who advances projects with its clients to transform ideas into concrete results, with pragmatism and high standards. The position: As part of the development of an artificial intelligence platform intended to securely exploit the internal knowledge of an international organization, the position aims to design, develop and operate an industrialized RAG (Retrieval-Augmented Generation) engine. The role covers the entire life cycle of the product, from the development of backend services to their operation in production. It also includes a strong agentic AI dimension, with the objective of making search and knowledge access capabilities directly usable by AI agents and by product teams. The environment is based in particular on Python, FastAPI, vector and hybrid search technologies, PostgreSQL, MongoDB, Kafka, Docker, Kubernetes, Terraform and Google Cloud Platform. The main missions will be as follows: Design, develop and operate the backend and API services of the RAG engine in Python and FastAPI, particularly with asynchronous processing.Build scalable ingestion pipelines to connect different business sources and process PDF, Office, HTML or scanned documents.Define and optimize strategies for parsing, chunking, embeddings and incremental content synchronization.Improve retrieval performance through vector and hybrid search, reranking, query rewriting and metadata filtering.Support search uses multilingual in an international context.Guarantee respect for access rights at the documentary level, data protection and auditability of processing.Measure and improve the quality of the RAG through indicators such as recall, precision, faithfulness and groundedness, as well as test sets and continuous evaluation mechanisms in production.Integrate access services to LLMs and embedding models, while optimizing costs and response times.Guarantee performance, scalability and reliability of the platform via monitoring, load testing and incident management, in an organization where the team ensures both the build and the run. Participate in build / buy / open source analyzes in collaboration with the architecture and product teams. Prepare the RAG engine for agentic uses by exposing it in the form of tools and MCP servers, while respecting access permissions. Implement agentic retrieval mechanisms, in particular the multi-step retrieval, query planning and tool-enriched search. Design SDKs, templates, plugins and documentation allowing product teams to quickly integrate their own knowledge bases. Use AI development agents on a daily basis to delegate all or part of the implementation, testing, refactoring, migrations and documentation, while retaining responsibility for the result. Formalize specifications and break development into tasks allowing efficient execution by multiple agents in parallel.Structure repositories suitable for agentic development through instruction files, reproducible environments, robust automated tests and CI guardrails.Share AI-assisted development best practices and help measure their impact on delivery speed and quality.Maintain a high level of quality through code reviews, automated tests, documentation and CI/CD pipelines under GitHub Actions and Google Cloud Platform.Required profile: The sought-after profile has at least 5 years of experience in backend development, with concrete experience on distributed systems and applications deployed in production. He has strong skills in Python, notably with FastAPI, asynchronous programming and API design. Real expertise in AI, NLP and information retrieval is expected. The profile includes the principles of embeddings, vector similarity, ranking and retrieval metrics, as well as how LLMs work. Above all, he knows how to evaluate and concretely improve the quality of a retrieval and general chain...

Apply here:

Web: Apply here

Emails: