AI Developer - Freelance (M/F)

France | Oct. 3, 2026

Report as Closed

Company: Collective.work

Country: France

Type: Onsite

Employment: Full-time

Description: Daily rate (TJM): 520

CONTEXT OF THE MISSION

The MTA (Modelling and Advanced Technologies) team designs and industrializes the AI ​​components used by the professions. Its main missions include:

- Development of APIs

- Development of chatbots: internal (e.g. Jarvis) and external (business assistants).

- Creation of AI agents: capable of executing workflows, interacting with APIs, orchestrating

tools, and automating complex tasks.

- Advanced integration of LLM models: Mistral, GPT, Claude, etc.

- Technical contribution to cross-functional platforms: IALAB, IAP, GED/PHX, RPA, etc.

- Participation in system demonstrations: chatbot subjects, automated workflows, IALAB, etc.

To support the rise of AI uses, the client wishes to strengthen the

MTA team with an experienced developer specializing in AI agents,

orchestration, advanced chatbots and IS integrations.

OBJECTIVES AND DELIVERABLES

Develop advanced AI agents for businesses

- Design of agents capable of calling APIs, triggering internal tools, and executing complex

actions.

- Multi-step scenario management: decision analysis action supervision.

- Construction of "AI tooling": actions, connectors, permissions, wrappers.

- Alignment with SSI, DPO and AI governance procedures.

2 Design or improvement of business chatbots

- Development of AI assistants to manage documentation, workflows, qualification

documentary, etc.

- Implementation of key mechanisms: memory, information retrieval,

contextualization.

- UX optimization of interactions: prompts, actions, structured responses.

- Multi-model support: Mistral Large, GPT-4o, Claude 3.7, etc.

3 Integration and orchestration with Bpifrance IS

- Secure calls to internal APIs: ERP, GED, RPA, CRM, PHX, etc.

- Connection to existing orchestration tools: RPA, IALAB, pipelines documents.

- Implementation of tool executors with fine management of authorizations and logs.

4 Industrialization and reliability

- Complete observability: logs, metrics, traceability, cost monitoring.

- Unit and end-to-end tests for chatbots and agents.

- Contribution to internal frameworks: in-house libs, conventions IA.

- Participation in system demos, PI planning and Agile increments.

Development

- Expert Python: FastAPI, async, modular architectures.

- Very good foundations in Node.js/TypeScript for certain front agents or integration services.

- API-first oriented development.

- Good practice of cloud architecture patterns: Azure Functions, containers, CI/CD.

- DévSecOps: You Build It And You Run It.

AI and Agents

- Mastery of the design of tooled AI agents (ReAct, Agentic Workflow, Toolformer-like).

- Knowledge of agent frameworks: LangChain, LlamaIndex, CrewAI, Semantic Kernel or

equivalents.

- Strong ability to model AI/business workflows in complex environments.

- Excellent mastery of LLM (Mistral, GPT, Claude).

Chatbots

- Design of complex conversational flows.

- Integration of memory modules, RAG, reformulation.

- Optimization of advanced prompts.

- Conversational UX awareness: turn-taking, clarifications automatic.

Integration and Data

- Internal REST/GraphQL API calls.

- Interaction with GED, documentary bases, data lakes.

- Use or construction of business connectors.

Quality, Security and Governance

- Logs, audit, monitoring: Grafana, Prometheus, Langfuse.

- Understanding security constraints agents: action permissions, order filtering, sandboxing, internal AI compliance.

Soft Skills

- Strong autonomy and ability to lead cross-functional topics.

- Pedagogy and popularization for professions using chatbots/agents.

- Ability to document: guides, API docs, diagrams.

- Product spirit: propose, challenge, prioritize, demonstrate.

-

Apply here:

Web: Apply here

Emails: