Working Student* (m/f/d) - Data Science & Deep Learning in the Context of Voice Assistant Systems - Karlsruhe or Remote

We are a startup for voice assistants and artificial intelligence from Karlsruhe, developing modern technologies for speech recognition and natural language assistant systems.

A focus of KENBUN is automatic speech recognition (Speech-to-Text) in industrial and medical environments. We already have very powerful models for German. Currently, we are further expanding our French speech recognition and have extensive training datasets with several million examples for this purpose.

Linguistic detail understanding is crucial for the quality of this Are the audio and transcription really correct? Are numbers spelled correctly? Are there systematic problems with homophones, contracted words, or other peculiarities of the French language? And which error patterns does our current speech recognition model show in practical use?

To support us, we are looking for a working student* with very good French skills and solid Python knowledge. The focus of the activity lies on linguistic quality control with existing datasets and models. You will work independently with our tools and data, but will be closely supported by our technical and scientific team.

If you are also interested in machine learning and speech-to-text models, you are welcome to delve into further topics such as evaluation or model training. However, this is not a prerequisite – your language intuition, analytical work style, and ability to systematically investigate data with Python are more important to us.

We look forward to your application!

Tasks

• Review of French audio and text data for linguistic and systematic errors
• Analysis of existing datasets with Python and our existing tools
• Identification and categorization of typical error patterns, e.g., with numbers, homophones, or contracted words
• Interactive testing of our French Speech-to-Text models through your own voice input
• Optional: Participation in evaluation, model training, or other machine learning topics

Qualifications

• French at a near-native level (C2 or comparable)
• German or English at least at B2 level
• Solid Python skills: You can understand existing scripts, make minor adjustments, and work independently with datasets
• Very good language intuition and enjoyment in systematically investigating linguistic peculiarities and errors
• Structured, meticulous, and independent work style
• Interest in speech recognition, language data, or Natural Language Processing is an advantage
• Machine learning experience is welcome but not a prerequisite
• Current place of residence in Germany and authorization to work in Germany

Benefits

• Work on real Speech-to-Text systems and large language datasets
• Insight into state-of-the-art methods from Speech Processing and Deep Learning
• Opportunity to learn from experienced machine learning and speech experts if interested
• Support in developing your potential through individual development
• Competent, individual support from our scientific and technical team
• Excellent working atmosphere in a highly motivated team
• Flexible activity, entirely possible remotely within Germany

Achieving great things together, courageously breaking new ground
and creating a work environment where everyone can unfold.

Do you want to help shape the future and make a difference? Then apply now!

Your contact person: Dr.-Ing. Simon Ottenhaus

We look forward to your application.

Find more English Speaking Jobs in Germany on Arbeitnow