About the role
Tonic builds AI products used by healthcare professionals in their daily practice. Tonic AI is a CE-marked Class IIa medical device, in live clinical use across multiple European markets. As the AI Quality & Validation Specialist you own how we evaluate our AI products, and your sign-off is a mandatory gate for every production deployment. Our quality processes are certified and audited, so what you build has to hold up when an external assessor asks how we concluded a given version was safe to release. This is not a traditional QA role. AI products are probabilistic, so quality cannot be reduced to right-or-wrong tests. The core of the job is deciding what “good enough” means, making that decision measurable and reproducible. You will work closely with Product, Data & AI, Engineering and other teams to continuously improve our products.
What you will do
- Own the release gate. Define the quality criteria a version must meet to reach production and run the evaluation against them.
- Build and maintain the validation system. Design test sets, define KPIs and evaluators, link them to the state-of-the-art evaluation practices and keep them current as the product and its clinical scope evolve.
- Make validation reproducible. Maintain a traceable chain from requirement to test case to dataset version to result to release decision, so the same conclusion can be reached again months later by someone else.
- Monitor products in production. Track quality and perform post-market surveillance to inform internal stakeholders on key findings.
- Stand behind the process in audits. Work with our compliance lead, who owns the quality management system, to keep validation documentation audit-ready and to represent the evaluation process during internal and external audits.
- Scale evaluation without scaling costs linearly. Validation coverage grows with every new feature, model and market, and adding reviewers is not a viable answer. You will coordinate the human review work that sits with our clinical team and find ways to reduce the human load.
What we are looking for
- 2-4 years or more in a role where evaluating quality was the job itself: model or LLM evaluation, applied AI quality, validation, or an equivalent function.
- You have built an evaluation framework from scratch or have managed and evolved one.
- Clear technical writing for a regulatory or audit audience. Validation reports and records are a primary deliverable of this role.
- Comfortable working with data and interpreting quality and performance metrics.
- Accountability as the role carries real authority over what ships, and the judgement to use it well.
- Strong communication skills and the ability to work effectively with AI engineers, clinicians and compliance.
- Collaborative mindset and ability to operate in a fast-paced environment.
- Full professional proficiency in Portuguese and English.
Nice to have
- Experience in a regulated environment, or with audited processes.
- Python or SQL as working tools. Useful, not required.
Why join us?
Healthcare professionals across four countries use our products in their daily work, and two more markets are opening. Tonic AI is a CE-marked medical device, which very few AI products in Europe can say, and it is used on real cases by real clinicians rather than sitting in a pilot that never leaves the slide deck.
If you want to work hands-on with AI that is already being used in real clinical practice and play a key role in making it more reliable, consistent and useful this is the job.
