(Senior) Machine Learning Engineer (m/w/d)
sipgate GmbH
About the role
About
sipgate develops communication software used by hundreds of thousands of people daily. We build AI not as a demo, but as a reliable part of our core products – from text and speech processing to intelligent assistant systems.
This role is for individuals who will own and be responsible for the entire model training pipeline: from data strategy, through training, to stable operation.
What We Do and Why We Need You
You will be responsible for the development, training, and operation of ML and LLM models – with a special focus on efficient, domain-specific models (Small & Medium Language Models) and clean ML/LLM Ops processes.
You will work at the intersection of model development, software engineering, and platform thinking.
Your Daily Work at sipgate Looks Like This:
Model & Training Work
- Develop, train, and fine-tune LLMs and SLMs (e.g., Mistral, Llama, domain-specific Transformers)
- Design efficient training strategies (LoRA/QLoRA, PEFT, Distillation, Transfer Learning)
- Build reproducible training pipelines, including dataset versioning
- Conduct structured model evaluation (offline metrics, task-specific benchmarks, regression tests)
ML / LLM Ops & Processes
- Build and further develop end-to-end ML & LLM Ops pipelines:
- Data ingestion & feature pipelines
- Training, fine-tuning & experiment tracking
- Model versioning, deployment & monitoring
- Establish process models for production-ready AI (similar to Software Development Lifecycles)
- Automate retraining, evaluation, and rollouts
- Monitor model drift, data quality, and inference performance
- Operate scalable inference services (batch & real-time)
System & Product Integration
- Develop RAG architectures with Vector Databases
- Close collaboration with Product & Backend to translate business requirements into model objectives
- Technical ownership of AI components in our products
- Code reviews, architectural decisions, and mentoring
What You Bring:
- You have (several years of) experience with the tools mentioned above
- You speak fluent German (at least C1)
- You live in Germany and are willing to collaborate with us in Düsseldorf at least once a month
How sipgate Works:
- We are agile from head to toe, and not just on paper. You can see how we work here.
- We have flat hierarchies and peer responsibility is our credo – this means: taking responsibility, making your own decisions, and looking out for each other are highly valued.
- We have a transparent salary model and no individual salary negotiations. You can find out exactly how this works here.
- Sustainability is important to us – our fleet consists only of electric cars, private cars can be charged for free during working hours, we get our electricity from solar power, we eat organic and regional food, and we try to avoid waste.
- We firmly believe in our 'Come as you are' principle and are on a journey to make sipgate as inclusive as possible, warmly welcoming people of all genders, ethnicities, body shapes, and social backgrounds to become part of our team.
... and if you want to see everything for yourself in person, feel free to stop by for lunch or breakfast (to the menu). #essenforfree
What Awaits You:
Would you like to learn more about what goes on behind the scenes? In our blog post, you can read about how we tick.
- Flexible working hours, no overtime, home office & 30 days of vacation
- Food & drinks are available in-house, on the house
- Free childcare during NRW school holidays & family events
- Subsidy for Urban Sports Club
- Free Deutschlandticket & subsidy for Jobrad
- Open Friday, unlimited training budget, free home office equipment & mobile phone for private use
Apply Now
Do you have questions or need more information? Feel free to contact Alena at E-Mail auf apply.j-vers.com ansehen (Job Number: #2493526).
Requirements
- Du hast (mehrjährige) Erfahrung mit den oben genannten Tools
- Du sprichst fließend Deutsch (mind. C1)
- Du wohnst in Deutschland und bist bereit mind. 1x pro Monat mit uns in Düsseldorf zusammenzuarbeiten
Responsibilities
- Du verantwortest den Aufbau, das Training und den Betrieb von ML- und LLM-Modellen – mit besonderem Fokus auf effiziente, domänenspezifische Modelle (Small & Medium Language Models) sowie saubere ML-/LLM-Ops-Prozesse.
- Du arbeitest an der Schnittstelle von Modellentwicklung, Software Engineering und Plattformdenken.
- Entwickeln, Trainieren und Fine-Tunen von LLMs und SLMs (z. B. Mistral, Llama, domänenspezifische Transformer)
- Design effizienter Training-Strategien (LoRA/QLoRA, PEFT, Distillation, Transfer Learning)
- Aufbau reproduzierbarer Trainingspipelines inkl. Dataset-Versionierung
- Durchführung strukturierter Modellevaluation (Offline-Metriken, Task-spezifische Benchmarks, Regressiontests)
- Aufbau und Weiterentwicklung von End-to-End ML- & LLM-Ops-Pipelines:
- Dateningestion & Feature Pipelines
- Training, Fine-Tuning & Experiment Tracking
- Modellversionierung, Deployment & Monitoring
- Etablieren von Prozessmodellen für produktionsreife KI (ähnlich Software-Development-Lifecycles)
- Automatisierung von Retraining, Evaluation und Rollouts
- Monitoring von Modell-Drift, Datenqualität und Inferenz-Performance
- Betrieb skalierbarer Inference-Services (Batch & Real-Time)
- Entwicklung von RAG-Architekturen mit Vector Databases
- Enge Zusammenarbeit mit Product & Backend zur Übersetzung fachlicher Anforderungen in Modellziele
- Technische Ownership über KI-Komponenten in unseren Produkten
- Code-Reviews, Architekturentscheidungen und Mentoring
Benefits
Skills
Don't send a generic resume
Paste this job description into Mimi and get a resume tailored to exactly what the hiring team is looking for.
Get started free