Senior Backend Engineer, AI Orchestration

Kestra Technologies; Deutschland; Remote; EUR 90.000 - 130.000; Vollzeit

Benefits dieser Stelle

Remote work

Health coverage

Home office setup

Zusammenfassung

Kestra is seeking a Full Stack Engineer to advance our Java-based orchestrator, focusing on AI workflow execution primitives and reliability. You’ll design core components, build plugins to connect LLM providers and agent frameworks, and collaborate with product teams to turn real-world use cases into reusable blueprints.

You will work in a remote-first environment with a modern tech stack (Java, Micronaut, Vue.js, Docker, Kubernetes) and opportunities to explore new tools as Kestra scales

Qualifikationen

• Proven experience building production-grade Java systems.
• Hands-on experience with LangChain4j or equivalent Java agent frameworks.
• Strong understanding of agentic workflows, retries, and observability.

Aufgaben

• Design and implement execution primitives for AI workflows inside the Kestra engine.
• Develop plugins connecting Kestra to LLM providers and agent frameworks.
• Translate real agent use cases into reusable blueprints with product teams.
• Ensure AI executions are observable and debuggable to the platform's standards.

Kenntnisse

Java

LangChain4j

Observability

Distributed systems

Tools

Micronaut

Vue.js

Docker

Kubernetes

Terraform

Kafka

Elasticsearch

PostgreSQL

MySQL

Jobbeschreibung

About Kestra

At Kestra, we’re on a mission to make orchestration and automation simpler for everyone. Our open-source platform helps teams manage complex workflows with confidence, and we’re already making a big impact in businesses around the world. Now, we’re looking for a Full Stack Engineer to help us take things to the next level.

In March 2026, we closed a $25M Series A led by RTP Global, with participation from Alven, ISAI, and Axeleo – backed by founders from Datadog, dbt Labs, and Hugging Face.

About the role

Kestra orchestrates workflows across systems and domains. AI agents and LLM-driven workflows are a growing part of that, and this role is about making sure they can run reliably inside the orchestrator. This is not a wrapper around a model API, but a platform orchestrating how agent steps should be run, retried, branched, and kept observable inside the orchestrator. That work happens in the Kestra core, which is Java.

What you would do

• Design and build the execution primitives for AI workflows in the Kestra engine

• Build and maintain the plugins that connect Kestra to LLM providers and agent frameworks

• Turn real agent use cases into reusable blueprints, together with the product team

• Keep AI executions observable and debuggable to the same standard as the rest of Kestra

What we are looking for

• A strong Java background. You have built and maintained production Java systems, not used Java in passing.

• Hands-on experience with LangChain4j or a comparable Java agent framework

• A clear mental model of how agentic systems are actually run: tool calls, state, retries, failure modes

• Comfort working close to an orchestration engine, where correctness and reliability matter more than demos

Our Tech Stack

• Backend: Java, Micronaut

• Frontend: Vue.js, Bootstrap

• Datastore: Kafka, Elasticsearch, PostgreSQL, MySQL

• Infrastructure: Docker, Kubernetes, Terraform

• Cloud: GCP, AWS, Azure

• Tools: GitHub for repo management, actions, and issues

We work with a variety of modern technologies, and you’ll get the opportunity to explore new tools and approaches as we grow.

Perks & Benefits

• Work from anywhere: We’re a remote-first company, so you can work from wherever feels like home. Plus, you’ll have access to coworking spaces worldwide if you ever need a change of scenery.

• Health coverage: From medical support, dental, and vision, we've got you covered.

• Home office setup on us: We’ll provide all the equipment you need to work comfortably.

Our Hiring Process

We aim to move quickly (2-3 weeks), but we can adjust the timeline if needed.

• Intro call with the hiring manager (30 min)

• Technical test (2 hour, homework done in async)

• Team chat with two of your future colleagues (30 min)

• Final discussion with one of our co-founders (30 min)

We’re here to make the process smooth and transparent for you.

Before you apply

This is a Java-first role. The AI ecosystem is mostly Python and we know that. We can work with someone who has built AI systems in Python (or in TypeScript via LangChain.js) and has a solid Java background to grow from. We cannot work with someone who only knows the Python AI stack and does not want to write Java.

We are also not looking for candidates whose AI experience is a single demo or a personal side project. We want people who have run AI in production, dealt with outages, and learned from the failures that come with it.