
Roche je globalno zdravstveno in farmacevtsko podjetje, ki se osredotoča na izboljšanje zdravstvene oskrbe pacientov z raziskavami in razvojem, digitalnimi inovacijami ter sodelovanjem. Z več kot 101.000 zaposlenimi združuje strokovno znanje s področij tehnologije, proizvodnje, poslovne strategije in drugih funkcij, ki podpirajo inovacije v zdravstvu. Roche poudarja raznolikost, enakost in vključenost v svojih mednarodnih ekipah ter ponuja karierne priložnosti za izkušene strokovnjake, študente in diplomante, vključno z možnostmi, povezanimi z njenim delovanjem v Nemčiji, Avstriji in Švici.
Senior Data Scientist, Agentic AI
Build production AI, generative AI, and agentic systems for Roche’s global healthcare business. Develop enterprise evaluation, governance, and MLOps capabilities.
170.100 zł – 315.900 zł / leto
HibridnoPolni delovni časSrednja ravenSeniorZahtevani Jezik: angleščinaPodatkovni znanstvenikOpis
- Move AI and machine learning solutions from prototype to monitored production.
- Build agentic AI systems that plan and carry out complex, multi-step work.
- Collaborate with product owners, engineers, and business stakeholders.
- Translate ambiguous problems into modelling approaches and explain their trade-offs.
- Choose suitable methods, from classical and fine-tuned models to foundation-model retrieval and agentic workflows.
- Write clean, reproducible code ready for production.
- Develop reusable patterns for the wider team.
- Set strong baselines, run held-out tests, and conduct detailed error analysis.
- Create evaluation datasets and test harnesses for generative AI systems.
- Use trusted datasets and prompt regression checks.
- Assess model outputs statistically for safety, reliability, and consistency, both offline and in production.
- Use real user feedback to improve solutions.
- Report AI cost, value, adoption, and usage analytics to leadership.
- Develop solutions within Roche’s governance, security, and compliance controls.
- Help build pipeline checks for fairness, explainability, and privacy.
- Share expertise through code reviews, demonstrations, and onboarding documentation.
Zahteve
- Practical expertise in data science and software engineering.
- Proficiency with Python, large language models, retrieval-augmented generation pipelines, and agentic workflows.
- Ability to write production code, design system architecture, and define integration standards.
- Experience evaluating and fine-tuning pre-trained models for domain-specific use.
- Experience designing retrieval-augmented generation and search-and-synthesis architectures.
- Experience building evaluation harnesses and statistically assessing outputs for safety, reliability, and consistency.
- Strong natural language processing skills, including preprocessing, language modelling, sentiment analysis, and named entity recognition.
- Strong grounding in statistics, model evaluation, feature engineering, and standard algorithms.
- Expertise in vector embeddings, search theory, vector database selection and architecture, prompt engineering, and context engineering.
- Experience building MLOps or LLMOps pipelines for continuous deployment and monitoring.
- Experience with LLM observability, including tracing, token usage, latency, cost telemetry, drift, and hallucination monitoring.
- Experience forecasting and optimizing infrastructure and API costs.
- Experience managing relationships with platform vendors.
- Strong written and spoken English.
- Ability to explain technical trade-offs to technical and non-technical audiences.
- Preferred: hands-on experience deploying enterprise AI solutions on AWS or Azure; familiarity with GCP is a plus.
- Preferred: knowledge of model risk management, fairness and bias mitigation, transparency, explainability, and data privacy.
- Preferred: experience in pharmaceuticals, healthcare, or financial services, including GxP compliance and handling sensitive information.
- Preferred: familiarity with emerging compliance standards for generative AI and foundation models.
Ugodnosti
- Annual performance bonus with a 15% target, based on performance.
- Dedicated budget for certifications, AI and technology courses, international conferences, and varied career paths.
- Two additional paid Fridays off each quarter.
- Sabbatical leave of up to three months.
- Option to work remotely from different global locations for specified periods during the year.
- Up to two additional paid weeks off for volunteering.
- Private healthcare through LuxMed packages.
- Group life insurance through UNUM.
- MultiSport card.
- Vacation subsidy.
- Roche stock purchase program.
- Annual sale of company laptops and cars.
- Hybrid work combining flexibility with in-person collaboration.
Sorodna delovna mesta
Ni najdenih delovnih mest
Poskusite prilagoditi ali počistiti filtre, da najdete več priložnosti.