Careers
Work on applied AI. Run with the team.
TW3 is a group of engineers working on applied and generative AI for companies and public institutions. We are looking for people who want to solve concrete problems, and who like to share more than an office.
Apply through the form, or at contact@tw3partners.com

Our teams
AI consulting and engineering
Designing and deploying AI systems for our clients, from agentic workflows to large-scale document intelligence.
Research
Sovereign generative AI (vision-language and vision-language-action models) applied to knowledge management and cognitive robotics.
Product
Building Citead.com, our platform for visibility in AI-generated answers.
Our culture
Small teams, short decision paths, direct access to the people who lead the group. Each engineer works on real client or R&D projects from the first weeks, with the responsibility that goes with it.
Sport at TW3
Sport is part of life at TW3. The company sponsors its engineers' participation in sporting events held in Île-de-France, such as marathons, trail runs and relays.
- Trail runs and relays
- Our engineers select races in Île-de-France, from short trails in the forests around Paris to longer formats, and put together TW3 teams.
- Charity races
- Some of the events support associations, which gives the team one more reason to take part.
- All levels welcome
- Formats are chosen so that everyone can join: short distances, relays, nordic walking. Supporters at the finish line are part of the team too.
- Voluntary
- Taking part is always a personal choice.
Other benefits
- Office
- In central Paris, in the heart of the Latin Quarter.
- Research and publications
- We encourage our engineers to write scientific articles, and support occasional participation in conferences.
- Working together on site
- Remote work is occasional. We favour working at the office or on client premises, because team dynamics are an important part of how we work.
Who we are looking for
- People who take ownership of a problem until it is solved.
- Engineers who can explain their work clearly, to a client as well as to a colleague.
- People who give and accept direct feedback.
How we hire
01
Introductory call with a member of the team.
02
Technical interview based on a problem close to our actual work.
03
Meeting with the founders to discuss the role and how we work.
Open positions (1)
Agentic AI EngineerTW3 Partners · Paris, 5th arrondissement · Fixed-term or permanent, on site
TW3 Partners is a Paris-based AI engineering and research firm. Its engineers and PhDs design and deploy AI agents for companies and public bodies across sectors (energy, pharma, industry, higher education), on a European cloud or on the clients' own servers, with open models. We are hiring an Agentic AI Engineer to take these agents from pilot to production, at scale and at a controlled cost.
The work
- Identify high-value use cases with business teams and turn them into agent specifications.
- Design and build the agents in Python: orchestration, tool calls, memory, RAG.
- Integrate the agents with the clients' existing applications and APIs.
- Select, adapt and serve models according to quality, latency and cost: fine-tuning, quantisation, batching.
- Build evaluation sets, measure quality before each release to production and put guardrails in place.
- Monitor agents in production (latency, errors, cost per user), in compliance with the GDPR and the AI Act.
- Run acceptance with the client, train the teams until they are autonomous and write the technical documentation.
Profile
- Engineering degree or master's in computer science, data science or artificial intelligence.
- Command of Python and of development practices: tests, Git, code review.
- Good knowledge of open models (Qwen, Llama, Mistral) and their ecosystem: Hugging Face, vLLM.
- Experience with agentic architectures and with evaluating LLM systems.
- Practice of production deployment (Docker, CI/CD) and of inference optimisation on GPUs.
- At ease with clients: gathering a need, restating it, explaining a technical choice.
- Autonomy, critical thinking and decisions backed by measurements.
- Open source contributions or personal projects around AI agents are a real plus.
Hiring process
- A 30-minute call or video call to present the role and your background.
- A take-home technical test in Python around a simple agent (tool calls, RAG, evaluation), followed by a debrief with the team.
- A technical interview on concrete use cases: the architecture of an agent for a client need, model choice and cost, scaling.
- A meeting with the team at our offices in the 5th arrondissement of Paris.
Conditions
- Contract
- Fixed-term or permanent depending on the profile. The position is not open to freelancers.
- Start
- As soon as possible.
- Location
- Paris, 5th arrondissement, on site.
- Salary
- 45k euros gross per year at hiring, moving up the salary scale as skills develop.
How we handle applicants' data is described in our privacy policy.