AI companies

Mistral AI careers: jobs, teams and how to get hired

Mistral AI has more open AI roles in Europe than any company on Tokn. Who it hires, where, what postings ask for and how it interviews.

Thomas Couderq, Founder of Tokn

8 min read

Tokn blog cover: Working at Mistral AI
In this article⌄

Mistral AI careers are mostly in Paris, but not only there. As of 12 October 2026, we track 84 open AI and technical roles at Mistral on Tokn, more than at any other company, and 42 of those postings list London among their cities. The biggest hiring block is not research: it is customer-facing deployment work.

Mistral tops our ranking of the AI companies hiring the most in Europe. Here is what working at Mistral AI involves: teams, cities, requirements, interviews and pay.

In short

  • 84 Mistral AI roles on Tokn as of 12 October 2026, 22 of them posted in the last 30 days.
  • About 43% are deployment and forward deployed roles that work with customers. Research is 17 roles.
  • Paris is the base, but half the postings also list London. Amsterdam and Munich follow.
  • Mistral is office-first: at least 3 days a week where stated. Only 4 roles are fully remote.
  • No salary appears on any posting. English is the working language; no AI posting requires French.

Mistral AI in 30 seconds

Mistral is a Paris AI lab that builds open-weight models and sells them, with tools and compute, to large companies.

FoundedApril 2023, Paris (Mistral's about page)
FoundersArthur Mensch (CEO), previously a researcher at Google DeepMind; Guillaume Lample (Chief Science Officer) and Timothée Lacroix (CTO), who came from Meta's AI research lab (SiliconANGLE)
SizeMore than 900 employees, according to its careers page
ProductsOpen-weight models, Le Chat (its assistant), Vibe (coding and work agent), AI Studio, Mistral Compute, Forge
Funding€3bn Series D in September 2026 at a post-money valuation above €21bn (Mistral's announcement)
Customers125+ enterprises in 20 countries, including Airbus, ASML and HSBC (same source)

Its newest model, Mistral Large 4, was released in preview on 6 October 2026. Open weights are promised by the end of October.

What Mistral AI is hiring for right now

Mostly people who put models to work at customers, then researchers, then infrastructure and product. The table sorts the 84 roles on Tokn into families by title.

Mistral's full careers board is larger; Tokn lists its AI and technical roles. On 12 October, Mistral's job board showed 215 postings worldwide, 150 based in Europe; the other European ones are mostly sales, HR, legal and finance.

FamilyRolesExamples
Research and science17AI Scientist, Research Engineer (Eval Platform, Forge, Inference Foundation), LLM Scientist (AI for Science)
Applied AI and forward deployed engineering16Forward Deployed ML Engineer, Technical Lead FDE, FDE Defense, Applied AI Engineer (Prototyping)
AI Deployment Strategist14Generalist in 7 cities, plus Science, Physics, Cybersecurity and sovereign institutions versions
Infrastructure, cloud and data10SRE Mistral Cloud, Network Automation, TPM HPC, Senior Data Engineer
Product8PM Vibe, PM Studio, Lead PM Inference, Product Director Cloud & Inference
Marketing and developer relations7Field Marketing (DACH), Technical Marketing Engineer, VP Global Developer Relations
Solutions architecture and operations6Infrastructure Solution Architect, Partner Solution Architect EMEA
Security and other6Offensive Security Engineer, Data Scientist, AI Engineer Product

Deployment and forward deployed roles

Deployment strategists, forward deployed and applied AI engineers and solutions roles add up to 36 of 84 (43%).

An AI Deployment Strategist scopes and sells the use case. Postings advertised in October asked for "2+ years in a client-facing strategic and technical role (e.g., data science consulting, value engineering, or technical sales)". A Forward Deployed ML Engineer builds it: "experience in Fine Tuning LLMs, tackling advanced RAG or agentic use cases", pre-sales calls included. The Applied AI team describes itself as "operating like startup CTOs who own end-to-end project execution".

Similar deployment roles in Paris, at Mistral and elsewhere:

Research and science

Research roles split into AI Scientists, who train and study models, and Research Engineers, who build the platforms around them: evaluation, inference, data infrastructure, Forge. A newer strand is AI for Science and Physics AI, grown after Mistral acquired Emmi AI, a Linz-based industrial simulation startup, in May 2026.

Infrastructure, product and the rest

Infra roles follow Mistral's push into its own compute and cloud. Product roles map to its products: Vibe, Studio, Context, Inference.

Where the jobs are: Paris, London, Munich and remote

Paris is home base, but many postings list several European cities. Two counts matter: the primary city, and every city listed.

CityPrimary cityListed on the posting
Paris6566
London642
Amsterdam320
Munich418
Warsaw111
Berlin08
Zurich15

Other cities listed on its postings include Madrid, Linz, Stockholm, Marseille, Lausanne and Luxembourg.

Mistral AI London jobs

42 of 84 postings list London among their cities, so half of Mistral's AI roles can be done from there. London is an office: postings refer to "our European offices (Paris, France and London, UK)". Our AI jobs in London page shows jobs by primary city only, so it lists about 6 Mistral roles, not 42. For all of them, check the locations on the Mistral AI jobs page on Tokn.

Munich and Linz

Mistral opened a Munich hub on 28 September 2026, with Physics AI and Industrial AI research teams and applied engineers working with partners such as BMW and Siemens Energy. Press reports put the German team at about 30 people, with around 100 planned by end-2027. Linz became an office through the Emmi AI deal.

Office policy and remote work

Mistral is office-first. Paris product and business postings say: "we encourage going to the office as much as we can (at least 3 days per week)", with each manager deciding how many remote days. Only 4 of the 84 roles are fully remote, all based in France.

One exception: some research engineering postings consider remote candidates in the countries they list, who visit the local office for their first week, then "at least 3 days per month".

Seniority, experience and internships

Where a minimum is stated (30 of 84, so treat these as examples), the median is about 5 years.

  • 2+ years: AI Deployment Strategist, Forward Deployed ML Engineer.
  • 4+ years "on large-scale ML codebases": Research Engineer, Machine Learning.
  • 5+ years: PM Vibe, SRE Mistral Cloud.
  • 8 years: Technical Lead FDE, Staff Deployment Strategist.

Junior roles are rare: 4 of 84, two of them Paris internships (see the FAQ).

What Mistral's job postings ask for

Python, LLM experience and comfort with customers. The skills we extracted most often:

SkillPostings
Python52
Kubernetes22
PyTorch18
AWS / GCP11 each
Azure, RAG, Go10 each
JAX9

Compare with the tools AI job postings ask for across the market. In the full text of the 84 postings, themes stand out:

  • Customers: 50 mention customers or clients, 26 mention pre-sales, 21 mention travel.
  • LLM work: 44 mention LLMs, 28 agents, 18 fine-tuning, 14 RAG or retrieval.
  • Open source: 30 mention it. AI Scientist postings ask for "publications or open-source projects".
  • Degrees: 11 mention a PhD, usually as "Master's or PhD in Computer Science (or equivalent proven track record)". A PhD is rarely a hard requirement.
  • Language: English is the working language. No AI posting on Tokn requires French; 3 mention German.
  • Clearance: 3 defence and sovereign roles ask for candidates who can obtain security clearance.

The Mistral AI hiring process, step by step

Mistral publishes its process. Its careers page describes two tracks, and 16 of the 84 postings list the steps.

  • Science, product and engineering: intro conversation, technical assessments ("typically 2–5 exercises"), a values conversation.
  • Go-to-market and corporate roles: intro conversation, 1 to 3 interviews on functional expertise and collaboration, a case study or role-based exercise, a values conversation.

"Reference checks are always conducted before an offer is extended." Mistral's postings describe these loops:

TrackSteps listed in postings
AI Deployment StrategistRecruiter screen; hiring manager (30 min); business acumen (45 min); technical deep dive on LLMs (60 min); panel with home prep; value talk (30 min); references
Data ScientistRecruiter (30 min); hiring manager (30 min); Python and stats/ML (60 min); 5-day homework and presentation (60 min); value talk (30 min); references
Software engineering (Mistral's board)Intro (30–45 min); live coding (45 min); system design (45 min); hiring manager (30 min); deep dive (60 min, leads and staff only); culture fit (30 min); references

Mistral AI salary and benefits

Mistral publishes no pay range on its postings: none of the 84 on Tokn, none of the 215 on its board. The only public figure is self-reported: on Levels.fyi, 22 Mistral software engineers in France report a median total compensation of €103,468 (base plus stock), as of 12 October 2026. That is a small sample for one role.

The careers page lists 100% employer-paid health premiums, 20 weeks of paid parental leave for birthing parents, childcare support, and relocation help including visa sponsorship. Benefits vary by country. For market context by role, see what AI roles pay in Europe.

How to stand out

The postings tell you what Mistral screens for. We'd focus on:

  1. Ship something with LLMs and show it. Fine-tuning, RAG and agents come up again and again.
  2. Contribute to open source. Postings name it as a plus, for research and applied roles alike.
  3. For deployment roles, prove you can face a client. Consulting, solutions engineering or sales engineering experience is asked for by name.
  4. Prepare for the case. Several loops include a home assignment and a presentation to defend.
  5. Read the values before the values talk. Mistral lists Audacity, Rigor, Customer centricity, Speed and Low-ego.
  6. Check every listed city. London, Amsterdam and Munich come up often.

For the wider method, read our step-by-step guide to landing an AI job or how to become a machine learning engineer.

Find Mistral AI jobs on Tokn

Mistral posted 22 roles in the last 30 days and 9 in the last week. See every current role on the Mistral AI jobs page, browse AI jobs in Paris or in London, and set a job alert to hear about new ones first.

FAQ

How many jobs does Mistral AI have open?

As of 12 October 2026, Tokn lists 84 Mistral AI roles in Europe, the most of any company we track. Mistral's own board had 215 postings worldwide, including sales, HR, legal and finance roles that Tokn doesn't list.

Can I work for Mistral AI from London?

Yes, for about half of its AI roles: 42 of the 84 postings on Tokn list London among their cities. Amsterdam and Munich come next.

Does Mistral AI hire remote workers?

Rarely. Only 4 of 84 roles are fully remote, all based in France, and postings that state a policy ask for at least 3 office days a week. A few research engineering roles accept remote hires in listed countries.

Do I need to speak French or have a PhD?

Usually not. English is the working language and no AI posting on Tokn requires French. Research postings accept a Master's "or equivalent proven track record" instead of a PhD.

What is the Mistral AI interview process?

Mistral's postings describe an intro call, a hiring manager interview, technical rounds or a case study, sometimes a panel or homework presentation, a values talk, then reference checks.

Does Mistral AI offer internships?

Yes, but few. Two were advertised in October, both in Paris: a 6-month Forward Deployed ML Engineer internship and an Applied Scientist internship. The first asked for a degree "from a tier 1 engineering school or university".

Open roles: AI jobs

All 826 jobs →

Keep reading

Get AI jobs in your inbox

Pick a role and a city: we email you new jobs as soon as companies post them. Free.

Create a job alert