Who we are
At OpenUp, our mission is to make mental well-being accessible to everyone. Following a ‘high tech & high touch’ approach, we support companies, their employees, and their families with proactive and practical mental health services. Through our digital platform, clients can access 1:1 check-ins, online courses, articles, videos, group sessions, mindfulness practices, and a team of committed mental well-being experts. As we believe that mental and physical well-being are deeply connected, we have also added Lifestyle Experts to our offering, focusing on nutrition, exercise, and sleep. We offer a truly holistic approach that helps people take ownership of their health. We’re proud to work with forward-thinking organizations and are committed to raising the bar on accessible care.
OpenUp was ranked in the Top 3 on LinkedIn’s Top Startups list for 2023, 2024 & 2025 and is continuously growing. We are expanding our international teams and we are searching for like-minded talent. We are already serving 2000+ companies (e.g. Rituals, Deloitte, Decathlon, KPMG, and McDonalds), through which 600K+ employees have access to our product.
About the role
As a Senior AI Engineer at OpenUp, you will own the design, build, and evolution of our AI-powered product experiences — starting with our AI Guide, which serves 600K+ users seeking mental and physical wellbeing support.
We're moving from single-prompt to multi-agent architectures that orchestrate tools, manage memory, and hold meaningful conversations at scale. You'll focus on conversational quality, evaluation infrastructure, and user trust — knowing the difference between a demo and a production system.
- Evaluation as a first-class concern. We operate in mental health, where a bad recommendation isn't just a UX issue. You'll build pipelines that measure conversational quality, safety, and clinical appropriateness — with regression detection before every deployment.
- Model-agnostic architecture. You'll design abstraction layers that allow seamless switching between Azure OpenAI, open-source models, and future providers.
- Prompt engineering at scale. You'll design, version, test, and optimise prompts across multiple agent personas — and build tooling that makes iteration fast, measurable, and safe.
You'll collaborate closely with Product, Design, Backend Engineers, and Security to translate product intent into scalable, trustworthy AI systems.
Responsibilities
- Multi-agent architecture and orchestration: Design and build agent-based systems that go beyond rigid scripts. Implement orchestration patterns including tool usage, inter-agent communication, memory management, and context handling. Own architectural decisions about when to use single-agent vs. multi-agent approaches and how agents hand off to each other.
- Evaluation and quality infrastructure: Build and maintain evaluation pipelines that catch regressions before users do. Design golden datasets, automated evaluation rubrics, and monitoring dashboards. Establish quality gates that every prompt change and model update must pass. This is the most critical part of the role in a mental health context.
- Prompt engineering and optimization at scale: Design, version, and optimize prompts across multiple agent personas and conversation types. Build the tooling and processes that make prompt iteration systematic: A/B testing frameworks, performance benchmarks, cost analysis, and safety checks. Treat prompt engineering as a proper engineering discipline with version control, testing, and rollback.
- RAG and grounding systems: Build AI systems grounded in OpenUp content using techniques such as RAG and agentic retrieval. Ensure the AI Guide draws from verified clinical content, articles, and expert knowledge rather than generating unsupported claims. Optimize retrieval quality, relevance, and latency.
- Production AI systems: Productionize AI solutions on Azure, including monitoring, observability, and continuous improvement. Own the operational health of AI features: latency, cost, error rates, and user satisfaction metrics. Work with the platform team on deployment, scaling, and reliability.
- Model-agnostic design for data sovereignty: Contribute to architecture decisions that keep the application layer decoupled from any single model provider. Design configuration-driven model selection and data isolation patterns that enable future private cloud and on-premise deployment scenarios for enterprise customers.
- Cross-functional collaboration: Collaborate with Product and Design to align AI behavior with intended user experience. Work with clinical advisors on safety and appropriateness. Balance user experience with privacy, safety, and compliance requirements (GDPR, EU AI Act).
About you
You are curious about how technology is rapidly evolving and our place in it, and motivated by building solutions that have a real impact on people's lives. You believe that learning comes from action, iteration and feedback, and that strong user value and trust go hand in hand.
- 5+ years software engineering experience, with at least 2+ years building and shipping AI or LLM-based systems in production
- Strong background in data science or machine learning. You understand statistical evaluation, experimental design, and can reason about model performance beyond accuracy scores
- Fluent in Python. Experience with .NET is a plus (our backend stack includes .NET, SQL Server, and ReactJS)
- Hands-on experience with multi-agent or agent-based architectures: orchestration, tool calling, memory, state management
- Experience designing and operating evaluation pipelines for LLMs or conversational systems. You have opinions on how to measure conversational quality and can build the infrastructure to do it
- Experience with prompt engineering at production scale: versioning, A/B testing, regression detection, cost optimization
- Working knowledge of Azure OpenAI or similar LLM platforms (OpenAI, Anthropic, Google)
- Familiarity with embeddings, vector search, RAG, and grounding techniques
- Ability to reason about AI risks: hallucinations, failure modes, adversarial inputs, and trust concerns in sensitive domains
- Bachelor's or Master's degree in computer science, data science, machine learning, or equivalent experience
- Fluent in English
Nice to have
- Experience with open-source LLMs (Llama, Mistral, or similar) and self-hosted model deployment
- Experience working in regulated or trust-sensitive domains (healthcare, mental health, fintech)
- Background in NLP, conversational AI, or dialogue systems
- Experience with Azure AI Agent Service, Semantic Kernel, or similar orchestration frameworks
- Familiarity with fine-tuning techniques and model adaptation
- Knowledge of EU AI Act compliance requirements
What you can expect
Joining OpenUp means being part of a fast-growing company that's doing something meaningful. You'll work with a diverse and international team. We like to have fun together with a lot of team building activities like our weekly ‘borrels’ and after work sport classes, the yearly ski-trip and our Padel tournament. Which makes our work fun, challenging and dynamic. Apart from this, we offer:
- €1500 annual personal development budget
- ClassPass: OpenUp contributes towards your classpass membership
- Monthly home internet/phone contribution (50EUR)
- Daily healthy and vegetarian lunch at our office.
- Regular office yoga, meditation classes or breathing exercises
- Flexible work model (hybrid and options for remote work).
- Monthly drinks and company wide summer fest and other fun activities.
- 27.5 vacation days per year.
- Free and unlimited access for you and your family to all our products, including our psychologists.
Next steps
Excited about this role at OpenUp? We love to learn more about you. Send us your application through our website and we will be in touch with you soon.
At OpenUp, everyone is welcome to apply - regardless of color, age, gender, nationality or any other label. We embrace what makes you unique and we believe that a diverse team is crucial to success. If you are ready to OpenUp, join us!
We may use artificial intelligence (AI) tools to streamline parts of our recruitment process. These tools assist our recruitment team but do not replace human judgement. All hiring decisions are made by humans. If you would like more information about how we use AI in the process, please contact us.
