AI Engineer for Speech / Natural Language Processing

Key Responsibilities

• Integrate third-party speech-to-text APIs (Deepgram, Whisper API, AssemblyAI, Azure, Google, etc.) with backend systems.
• Evaluate and fine-tune STT models for European languages, especially Swiss-German and German dialects, High German, and English.
• Build backend services (Java/Python/REST API) to process transcripts, handle streaming audio, and manage API calls.
• Implement NLP pipelines for:
• Named-entity extraction (name, address, phone numbers, etc.)
• Parsing long free-form sentences into structured fields
• Confidence scoring, error handling, and validation
• Optimize performance for noisy audio, accents, and dialect variations.
• Work with Azure/AWS/GCP for deployment, monitoring, logs, and scaling.
• Collaborate with product team to integrate transcription + NLP into form-filling workflows.

Required Skills

• Experience integrating speech-to-text APIs; familiarity with Whisper / Deepgram is a plus.
• Strong programming skills in Java or Python, and ability to build/consume REST APIs.
• Practical knowledge of NLP techniques (NER, regex, text classification, entity extraction).
• Experience with cloud platforms (Azure preferred).
• Ability to evaluate model accuracy for Swiss-German dialects etc and improve results with preprocessing/post-processing.
• Understanding of streaming audio, audio formats, and latency constraints.
• Good debugging skills and ability to handle ambiguous, noisy, or accented speech input.

Desired Skills

• Exposure to fine-tuning Whisper or similar models for dialect speech.
• Familiarity with German/Swiss-German linguistic variations.
• Experience with Docker, cloud functions, or scalable backend design.

Job Type: Freelance

Work Location: Remote