Senior Machine Learning Engineer (ASR) (all)

Ai-enhanced
Berlin, Germany
4 days ago

Role details

Contract type
Permanent contract
Employment type
Full-time (> 32 hours)
Working hours
Regular working hours
Languages
English, German
Experience level
Senior

Job location

Remote
Berlin, Germany

Tech stack

Audio Signal Processing
Computer Programming
Python
Machine Learning
TensorFlow
Speech Recognition
PyTorch
Deep Learning
Kubernetes
Information Technology
Production Code
Machine Learning Operations
Data Pipelines
Docker

Job description

  • ASR Model Development: Fine-tune, evaluate, and deploy state-of-the-art models in speech recognition and audio processing (e.g., Whisper, wav2vec).
  • Data Curation & Annotation: Collect and curate custom ASR datasets, including data sourcing, annotation pipeline setup, quality control, and alignment/segmentation procedures.
  • Audio Pipelines: Build and maintain robust data pipelines and audio preprocessing workflows for clinical environments.
  • MLOps Collaboration: Work closely with the MLOps team to ensure continuous training, monitoring, and seamless deployment of models in production.
  • Experimentation: Design and conduct experiments to validate new approaches, datasets, and architectures to improve accuracy in noisy or specialized medical settings.
  • Cross-Functional Impact: Collaborate with product managers and developers to translate complex speech solutions into production-ready healthcare tools.

Requirements

  • Education: Master's degree in Computer Science or Engineering.
  • Experience: Typically 5-8+ years of experience in ML engineering.
  • Technical Proficiency: Strong programming skills in Python (working with production code and deploying models in production) and ML frameworks (PyTorch, TensorFlow or Jax).
  • ASR Expertise: Direct experience with ASR models (Whisper, wav2vec, …), VAD, alignment and diarisation, and complex speech/audio processing pipelines.
  • Deep Learning: Extensive experience with transformer-based architectures and deep learning models.
  • MLOps Foundation: Practical experience with MLOps pipeline components such as Docker, MLflow, W&B, DVC, or Kubernetes.

Nice to Have

  • Domain Knowledge: Knowledge of multilingual or domain-specific modeling (specifically medical, legal, or other specialized terminologies).
  • Scalability: Experience with distributed training systems for large-scale model optimization.
  • Language Skills: Full proficiency in English and German.

Benefits & conditions

  • Full ownership of impactful ML components in a fast-growing AI environment.
  • A collaborative team that values curiosity, learning, and pragmatic problem-solving.
  • Flexible working arrangements (remote or hybrid).
  • 30 days of annual vacation.
  • Competitive salary depending on experience.
  • The chance to work on products that directly shape the future of healthcare.

About the company

With Sonia, doctors are successful doctors. We create and deploy AI-enhanced solutions that make doctors' lives easier, patients' care better, and healthcare systems more efficient. If you're an intrinsically motivated self-starter who values impactful work, join us in revolutionizing healthcare., At Sonia, we're building more than just a product - we're building a team that helps doctors succeed with the power of AI. To grow and shape the future of healthcare, we're looking for people who want to create, challenge, and collaborate - and have fun while doing it. We believe teams thrive when different perspectives come together and work toward a shared goal. We welcome and respect everyone for who they are, regardless of origin, gender, age, religion, disability, or sexual identity. We're excited to meet people from all backgrounds who want to bring their unique perspective to our mission. Are you in?

Apply for this position