AI team
ML Engineer
Zurich, Switzerland · Hybrid · Full-time
Join the mission to make communications trustworthy in the age of AI
Own audio ML end to end: turn messy real-world recordings into detection models, prove they work, and get them running in production against voice fraud that improves every few months.
About us
Aurigin.ai is a Zurich-born startup on a mission to restore trust in digital communication by protecting high-stakes organizations from AI-generated voice fraud in real time. Our deepfake detection is used to stop account takeovers, executive impersonation, and fabricated recordings before they cause damage.
We are a team of engineers, researchers, and builders with backgrounds across tech, consulting, and startups. We thrive on collaboration, embrace a "move fast to launch and iterate" mentality, and share a vision of a safer, more transparent world in the age of AI.
Why we need you
Our real-time detection engine is in production with its first clients, and the attacks it faces get better every few months. Keeping ahead of them is a full-time job for someone who is comfortable with terabytes of noisy audio, rigorous about evaluation, and able to take a promising experiment all the way to something that runs reliably on live traffic. Today that work is spread across the team. We want one person who owns it.
What you will do
- Handle audio data at scale: design and implement robust data pipelines and feature-extraction modules for large-scale audio datasets
- Bring models to life: prototype, evaluate, and iterate on machine learning models for deepfake detection and language-related tasks
- Measure what matters: define performance metrics, design experiments, tune hyperparameters, and validate models with rigorous statistical methods
- Ship it and keep it working: take models to production, watch how they behave on real traffic, and close the gap when lab results and live results disagree
- Stay ahead of the attack: follow new speech synthesis and voice conversion methods, and test our detectors against them before someone else does
About you
- Hands-on machine learning work with audio or speech, from a job, research, or your own projects. We care about what you have built, not how many years it took
- Technical proficiency in Python, PyTorch, and Hugging Face tools, with familiarity in audio processing libraries (e.g. librosa, torchaudio)
- Strong grasp of audio engineering principles and best practices in feature engineering
- Ownership mindset: self-driven, love solving complex problems, and can run projects end-to-end in a fast-paced startup
- Fluent in English (any other language is a plus)
- Hybrid in Zurich (at least two days a week in the office)
Nice to have
- Experience with Hugging Face audio models, including fine-tuning for classification or sequence modeling
- Knowledge of audio deepfake detection, voice conversion, or speech synthesis
- Contributions to open-source ML or audio projects
- Real-time audio or streaming data experience
Benefits
- Work that ships: your models run against live fraud instead of ending up in a report
- Help shape the product and the team from an early stage
- A small enough team that you are in the room for the decisions that affect your work
- Flexible hybrid setup: work where you are most productive, with at least two days a week in the Zurich office
- CHF 90k salary plus ~CHF 10k in stock options, so you hold a share of what you build
Ready to apply?
If you want to use your skills for good and help make communications trustworthy, send your CV and a short note to ai-careers@aurigin.ai.
Apply via email