Missions
- Clear the failure state, and prevent the next one. Diagnose problems across Prefect and Dagster installations, integrations, deployments, and the cloud and container environments they run in. Find the underlying "why" so your first response resolves the question and pre-empts the follow-ups. Turn what you learn into docs, debugging resources, and tooling that deflect repeat submissions.
- Own what's ours; say no clearly to what isn't. Take full accountability for root causes inside our products' domain of control. For problems outside our Scope of Support, issue a principled no that still gives the customer a way forward. Helplessness erodes trust as fast as a wrong answer.
- Support your findings, and ask customers to support theirs. Back every diagnosis with evidence and logic. Assume the customer is the expert on their own stack, ask for what you need to troubleshoot both sides of an issue, and explain why you need it. Leave room for the possibility you've missed something.
- Give the best possible update every time, and say so when you're still building it. A half-baked answer sent fast is worse than a complete one sent when it's ready; but radio silence is worse than both. When an issue needs to be escalated, you stay the customer's single point of contact, keep them updated on a regular cadence, and close the loop.
- Feed the product. File reproducible product issues and surface recurring failure patterns to Engineering and Product.
You'll work the queue independently on day-to-day issues and own moderate-complexity problems end to end, including investigations across a customer's cloud environment. You'll ask when the Scope of Support or an escalation path is unclear, and you'll build the judgment to recognize when an issue belongs with Customer Engineering or a CSM instead. We measure the role on Customer Effort Score and resolution time, held in tension: a fast answer that costs the customer effort isn't a good one.
Timezone & on‑call: Available to work 09:00 – 18:00 CET (UTC ± 2) and participate in the 24/7 on‑call rotations.
Profil recherché
- Python fluency at the level of reading and debugging someone else's code, not only writing your own.
- Hands-on experience running workloads in at least one major cloud and in containers or Kubernetes, enough to troubleshoot a deployment you didn't build.
- Experience with a workflow orchestrator (Prefect, Dagster, Airflow, or comparable) and a strong interest in the data engineering space it lives in.
- A track record of technical, customer-facing work: support, solutions engineering, developer relations, or similar.
- Clear enough written communication to explain a root cause to a skeptical senior engineer and earn their trust.
Description du poste
PREFECT https://www.prefect.io/ BUILDS AND OPERATES RESILIENT, PYTHONIC ORCHESTRATION AND MCP PLATFORMS -- PREFECT OSS https://github.com/PrefectHQ/prefect, PREFECT CLOUD https://www.prefect.io/prefect/cloud, FASTMCP OSS, https://gofastmcp.com/getting-started/welcome AND HORIZON https://www.prefect.io/horizon-- USED FOR MISSION-CRITICAL WORKLOADS.
Our Vision: Prefect will define automation for the context era.
Our Mission: Curate an intelligent context layer that delivers the right information at the right time.
About Prefect
Prefect builds automation for an unpredictable world. The last decade of automation was about protecting workflows from unpredictability; in the agentic era, unpredictability is the point. Our mission is to give people confidence in automated work, whether that work is a mission-critical data pipeline or a fleet of AI agents.
In August 2026, Prefect and Dagster, competitors for eight years, joined forces to build the next generation of automation infrastructure. Today, our ecosystem includes Dagster, the data platform; Prefect, the agent platform; and FastMCP, the leading open-source developer framework for the MCP ecosystem. Together, our open-source and commercial products are trusted by Fortune 500 companies, data innovators, and high-growth technology companies around the world.
We think of our culture as the operating system of the company. We’ve carefully created a supportive, high-performance environment that empowers our team to do the best work of their careers, have meaningful impact, and continue growing personally and professionally. We’re a remote-first team that values high standards, ownership, and thoughtful collaboration.
ABOUT PREFECT
Prefect builds automation for an unpredictable world. The last decade of automation was about protecting workflows from unpredictability; in the agentic era, unpredictability is the point. Our mission is to give people confidence in automated work, whether that work is a mission-critical data pipeline or a fleet of AI agents.
In August 2026, Prefect and Dagster, competitors for eight years, joined forces to build the next generation of automation infrastructure. Today, our ecosystem includes Dagster, the data platform; Prefect, the agent platform; and FastMCP, the leading open-source developer framework for the MCP ecosystem. Together, our open-source and commercial products are trusted by Fortune 500 companies, data innovators, and high-growth technology companies around the world.
We think of our culture as the operating system of the company. We’ve carefully created a supportive, high-performance environment that empowers our team to do the best work of their careers, have meaningful impact, and continue growing personally and professionally. We’re a remote-first team that values high standards, ownership, and thoughtful collaboration.
ROLE SUMMARY
Developer Success is the technical support function for Prefect's customers across the Prefect and Dagster product lines. Its mission is to resolve customer technical issues with the least possible customer effort, and to prevent those issues from recurring.
We are hiring a Developer Success Engineer to work the combined Prefect and Dagster support queue. You will diagnose and resolve technical issues in customer environments, own each issue through resolution, and contribute to the documentation, tooling, and product feedback that reduce future volume. You will be trained on both Prefect’s and Dagster’s products.
The team is investing in agent-based automation for routine support work, which is moving the engineering focus toward complex diagnosis and escalation. This role is well suited to an engineer who wants to build deep expertise in orchestration and data platform environments and apply it directly with customers.
WHAT YOU'LL DO
- Clear the failure state, and prevent the next one. Diagnose problems across Prefect and Dagster installations, integrations, deployments, and the cloud and container environments they run in. Find the underlying "why" so your first response resolves the question and pre-empts the follow-ups. Turn what you learn into docs, debugging resources, and tooling that deflect repeat submissions.
- Own what's ours; say no clearly to what isn't. Take full accountability for root causes inside our products' domain of control. For problems outside our Scope of Support, issue a principled no that still gives the customer a way forward. Helplessness erodes trust as fast as a wrong answer.
- Support your findings, and ask customers to support theirs. Back every diagnosis with evidence and logic. Assume the customer is the expert on their own stack, ask for what you need to troubleshoot both sides of an issue, and explain why you need it. Leave room for the possibility you've missed something.
- Give the best possible update every time, and say so when you're still building it. A half-baked answer sent fast is worse than a complete one sent when it's ready; but radio silence is worse than both. When an issue needs to be escalated, you stay the customer's single point of contact, keep them updated on a regular cadence, and close the loop.
- Feed the product. File reproducible product issues and surface recurring failure patterns to Engineering and Product.
You'll work the queue independently on day-to-day issues and own moderate-complexity problems end to end, including investigations across a customer's cloud environment. You'll ask when the Scope of Support or an escalation path is unclear, and you'll build the judgment to recognize when an issue belongs with Customer Engineering or a CSM instead. We measure the role on Customer Effort Score and resolution time, held in tension: a fast answer that costs the customer effort isn't a good one.
Timezone & on‑call: Available to work 09:00 – 18:00 CET (UTC ± 2) and participate in the 24/7 on‑call rotations.
YOUR QUALIFICATIONS
Must have:
- Python fluency at the
GetGlobalJob n'est ni l'employeur ni un intermédiaire de recrutement. La candidature se fait sur le site d'origine : vérifiez toujours l'annonce avant d'envoyer vos données et ne payez jamais pour un emploi.