Founding Field Engineer, France

The first field engineer we hire in France, joining a Field Engineering team of six worldwide. Elsewhere this job is called Solutions Engineer, Solutions Architect or Pre-Sales Engineer. The level follows your background: Senior or Principal.

Location: Remote, from anywhere in France Contract: Full-time, CDI Team: Field Engineering, 6 engineers worldwide, reporting to our Lead Field Engineer Languages: French and English, both fluent

KESTRA

Kestra is an open-source orchestration platform: one control plane for everything that runs inside a company, from data pipelines and infrastructure automation to business processes and, now, AI agents.

We were born in France and the engineering is still here. Our growth now runs through the United States, where an increasing share of our new enterprise business is won, while the French business expands at the same time. In March 2026 we closed a $25M Series A led by RTP Global with Alven, ISAI and Axeleo, alongside founders from Datadog, dbt Labs and Hugging Face. Both sides are accelerating, and this job is part of that.

What we replace. Control-M. VMware vRA and Aria. Automic. Airflow. Dagster. Prefect. Camunda. AWS Step Functions. Cron at industrial scale.

Where Kestra runs. Teams at JPMorgan Chase, Apple, TotalEnergies, Crédit Agricole, Leroy Merlin, Bloomberg, Banque Française Mutualiste and T-Systems run Kestra in production, alongside around a hundred other very large groups.

WHERE THIS IS GOING

Every large company today runs four orchestration stacks that do not talk to each other. One for data. One for infrastructure. One for business processes. One for the batch chains inherited from the mainframe era. Each was bought by a different team in a different decade, each has its own scheduler, its own audit trail, its own way of failing at 3am. Nobody in the building can answer the question "what is running right now, and what stops if this stops".

Agents are about to make that worse, fast. They are being deployed exactly the way data pipelines were deployed in 2015: one team at a time, each with its own runner, no shared governance, no common view. An autonomous system nobody can see is not something a bank puts into production, and every CIO we speak to already knows it.

Our bet is that those four stacks collapse into one control plane, and that whoever gets there first defines the category for the next ten years.

Kestra orchestrates up to 8 million workflow executions a day, and that counts only the instances reporting telemetry back to us. The number matters less than what sits behind it. A bank's payment chain. The provisioning of thousands of servers. The nightly close that has to finish before the markets open. The pipeline a factory plans its week on. Once a company puts all of that on one control plane, the orchestrator stops being a tool inside one team's stack. It becomes the one place where anyone can answer what is running right now, who authorised it, and what stops if it stops.

That is what we are going after. Not a better scheduler. The layer every large company is going to need, and will install only once.

WHY THIS JOB EXISTS

Our French accounts are among our largest and most demanding: banks, insurers, industrial groups, energy, retail. These are not pilots on a side project. They are orchestration migrations that touch thousands of servers and the processing chains the business depends on.

The engineering is rarely the hard part. The hard part is the six months between "this looks promising" and "our production runs on it": mapping a setup nobody fully documented, agreeing in writing what success means, convincing the architect who was perfectly happy with the old tool, becoming the person a CTO calls first.

We are opening this job in France because those accounts deserve an engineer in their timezone, in their language, who can be in the room.

WHAT YOU WILL DO

Before the contract

- Technical discovery with the Account Executive: what runs today, what it costs, what breaks, who owns it, and who inside the company would rather nothing changed.

- You own the evaluation. Our POCs are deliberately tight: one scoping session, 2 or 3 use cases with success criteria written down, self-hosted, no data leaving the customer's environment, 2 to 3 weeks start to finish. You write the success plan and you answer for the verdict.

- You build the demo. Not a slide deck of one: a running environment, their data shape, their constraints.

- You design the target architecture and defend it in front of architects who have already picked their favourite.

- You make the case against the tool in place. We are not subtle about this: when a prospect runs Control-M, vRA or a fleet of Airflow instances, we say what we replace and why.

After the contract

- Onboarding and training. The onboarding and follow-up framework is this team's own work, built over the past year. You use it and you sharpen it.

- Escalations on your accounts: diagnose with our core engineers, keep the customer informed while the situation is unresolved, then document the cause and build the automated check that prevents the next one.

- Quarterly architecture reviews with the people who operate Kestra day to day.

- The next use case. Most of our growth comes from a fifth team inside an existing customer discovering they can put their own processing on the platform.

- The voice of the field inside Product. Our PMs prioritise on what you bring back, so bring it back with evidence.

With the community and the product

- Answer open-source users on Slack and GitHub. Several of our largest customers started as someone asking a question there.

- Write what outlives the call: blueprints, how-to guides, migration notes, short videos.

- Speak at meetups and webinars if you want to. An option, not an obligation.

WHAT YOU TOUCH EVERY WEEK

Kubernetes, Terraform or OpenTofu, Docker, YAML (our flows are declarative YAML), Git and GitHub, CI/CD pipelines, and the security basics of an enterprise deployment: RBAC, secrets, network isolation, air-gapped installs. Plus whichever of Python, SQL, dbt, Spark or Kafka the account in front of you runs on.

WHAT YOUR FIRST YEAR LOOKS LIKE

Month 1. Kestra Enterprise running on your own cluster, one full migration done end to end in a lab, and you have sat in on discovery calls and at least one escalation.

Month 3. Two or three accounts of your own, and your first evaluation run solo.

Month 12. A portfolio of French accounts, some of them among our largest. Account teams route architecture questions to you by default, and at least one chapter of the Field Engineering playbook has your name on it.

WHAT WE ARE LOOKING FOR

This is a senior job. We do not train the technical half here, and we will not put someone in front of a CAC 40 architecture committee for the first time.

Required

- A high, verifiable technical level. You write production code, you debug someone else's, and you know what happens inside a Kubernetes cluster when it goes wrong. Operations, not demonstrations.

- Time in a startup or scale-up funded at Series A, B or C. Not a prestige criterion: in those companies nobody catches things behind you. You have held a scope alone, with no support team underneath and no process written in advance. That is exactly this job.

- Long tenures. At least two roles of 2 years or more, in companies whose systems have measurable size. A string of twelve-month stops tells us nothing about your ability to finish what you start.

- Complex projects you owned yourself, not ones you attended. A migration, a rebuild, a critical go-live: you can walk through the scope, the constraints, what broke, and what you decided at that moment.

- You have been alone in front of a customer. Workshop, POC, technical committee, incident. This job is as much consultant as engineer, and the consultant half cannot be improvised.

- Depth in at least one of: data engineering (production pipelines, dbt, Spark, Snowflake, DuckDB); platform and IaC (Terraform or OpenTofu, Kubernetes, CI/CD, and the internal politics of a platform team); AI and MLOps (LLM lifecycle, vector databases, models out of notebooks into production, agent orchestration under governance).

- You have operated or replaced an orchestrator: Airflow, Prefect, Dagster, Control-M, Automic, vRA, Step Functions, or cron at industrial scale. You know the subject from the inside, not from a comparison table.

- Flawless French, and English good enough to argue about architecture with engineers spread across several countries.

- You explain the same architecture twice: once to a CTO, once to a junior analyst, and both walk away with it.

Appreciated

- Pre-sales or field experience at a software vendor.

- Regulated or constrained environments: banking, insurance, energy, defence. Air-gapped deployments, EU data sovereignty, RBAC and audit as opening questions rather than afterthoughts.

- A public trace: open-source contributions, talks, technical writing.

- A third language, German above all.

WHY LEAVE A LARGE VENDOR FOR THIS

In pre-sales at a company of several thousand people, you know the ceiling: you prove the product, then you hand the account over, and the architecture you argued for is executed by someone else. Here the same person runs the evaluation, designs the target architecture, and stays on the account for the three years that follow. Six field engineers cover the world, so what you decide becomes the method rather than a local exception. And the product is open source: you can read the code you sell, open an issue, and watch it get fixed by someone you speak to every week.

THE HARD PARTS

We would rather you read this now than find out in month three.

- You will sometimes be the only technical person a customer trusts during an incident.

- Some evaluations are lost, and some are lost for reasons that have nothing to do with the product.

- Enterprise migrations are slow. If you need a visible result every two weeks, this job will wear you down.

- The Field Engineering playbook is half written. You will write parts of it while doing the job, not before.

WHAT WE OFFER

- A great compensation, plus BSPCE.

- Remote-first, flexible hours, no core-hours theatre.

- A whole-company offsite every year.

- Equipment budget.

- Conference and training budget, and the time to use it.

HOW WE HIRE

1. 45 minutes with our Lead Field Engineer: your background, one architecture you have operated end to end, and your questions.

2. 45 minutes with Emmanuel, our CEO.

HOW TO APPLY

Send your CV and one page on an orchestration setup you have operated or replaced: what it ran, what broke when it grew, and what you would rebuild differently today. We read the page first.