Applied Data Scientist

We know you have access to GPT. We use them too. But for your application, please leave the LLM off. We would much rather read two sentences of genuine, slightly imperfect English about who you are than a flawless, five-paragraph AI-generated essay that sounds like everyone else.

Spacebackend revolutionizes system integration in the aerospace industry through AI-driven interoperability across various hardware components. Our product Lynapse streamlines the entire satellite integration workflow by enabling modular integration across payloads, components, and subsystems. The Lynapse core technology transforms hardware specifications into digital twins and automatically generates, tests, and validates hardware-agnostic, OS-agnostic on-board satellite middleware software. Lynapse is a web-based collaborative modeling and design tool for mission-critical projects, allowing teams to create, test, validate, and simulate data interfaces for dozens or hundreds of hardware components composing satellites, spacecraft, or lunar landers. 

As an Applied Data Scientist & AI, you will architect and implement advanced language and intelligence models to convert technical interface specifications into digital-twin representations, identify gaps or ambiguities in those specifications, and lay the groundwork for unified on-board spacecraft autonomy.

• Design and train custom transformer-based models to ingest manufacturer interface specifications and generate structured, machine-actionable digital-twin schemas. 
• Leverage embedding methods, retrieval-augmented generation, and knowledge-graph techniques to ensure high-fidelity mapping of hardware capabilities.

• Fine-tune existing large-scale models to flag missing or conflicting information in source specifications and generate curated feedback for engineers. 
• Develop metrics and evaluation pipelines to quantify specification coverage and ambiguity reduction over iterative training cycles.

• Prototype neural and hybrid architectures (e.g., CNNs, reinforcement learning policies, LLMs) that use the digital-twin representation to command and simulate spacecraft behaviors. 
• Collaborate with systems and embedded teams to define interfaces between model outputs and on-board flight software modules for autonomy.

• Lead hypothesis-driven research: design experiments, collect data, analyze results, and publish findings internally. 
• Establish reproducible ML infrastructure: experiment tracking, model versioning, and containerized training pipelines (e.g., MLflow, Docker, Kubernetes).

• Guide junior researchers and interns, foster knowledge sharing, and promote an open-minded, iterative R&D culture.

We are interested in every qualified candidate who is eligible to work in Luxembourg. However, we are not able to sponsor visas for this position. 

We are seeking a researcher who:

1. Bridges Theory & Practice: You translate creative AI research into robust, production-ready pipelines. 
2. Iterative & Results-Oriented: You prototype rapidly, validate against real datasets, and refine models in agile cycles without losing momentum. 
3. Collaborative Innovator: You collaborate across functions (embedded, systems, DevOps) to integrate AI outputs into broader satellite workflows. 
4. Aspires to Leadership: an experienced engineer excited to grow into a senior R&D manager, and to take on increasing responsibility and shape our research directions over time.
5. Has a startup Mindset: You are going to work in an early-stage startup, your state of mind is flexible, open to creativity, enthusiasm for solving complex things.

1. MSc or PhD in Computer Science, AI/ML, Electrical Engineering, or related field 
2. 3+ years of hands-on AI/ML/NLP/DNN/CNN research or engineering experience 
3. 2+ years of experience in writing Python code and algorithms. 
4. Expertise in transformer architecture, fine-tuning frameworks (e.g., Hugging Face Transformers) 
5. Strong Python proficiency and experience with PyTorch or TensorFlow 
6. Experience building data pipelines for large technical document collections 
7. Familiarity with reinforcement learning or neural control methods 
8. Practical knowledge of containerized, scalable AI/ML infrastructure (Docker, Kubernetes) 
9. Demonstrated record of published research or open-source contributions

1. Background in electronics, aerospace or robotics autonomy research 
2. Experience in knowledge graphs, ontologies, or graph neural networks 
3. Familiarity with simulation environments (Gazebo, CoppeliaSim) or digital-twin frameworks 
4. Prior work on uncertainty quantification or interpretability in ML models

1. Salary + stock options package 
2. Flexible hours & on-site work at Technoport Belval, Luxembourg 
3. Career path to technical lead or executive roles
4. Opportunity to pioneer AI for satellite autonomy 

We understand the reality of the current job market: you are competing against hundreds of "perfect" automated resumes, which forces you to polish every word and cast a wide net just to get noticed. We want to bypass that noise and respect your time. Our hiring process is straightforward, transparent, and entirely human. We start by reviewing your initial application for a high-level mutual fit. If we believe there is a match, the next step is simple: we will ask you to send over a short, 60-to-90-second video just talking to us - no production value needed, just you. From there, we move into live, collaborative conversations where we look at actual code together and discuss your technical approach. We do not use unpaid take-home projects or automated quizzes in our hiring process.

Please send your CV and links to relevant projects or GitHub repos to [email protected] with the subject line: Applied Data Scientist & AI

Important note: We know that no candidate meets every single requirement. However, if you’re passionate about the role and meet most of the requirements, please contact us!