Internship - Evaluation of Retrieval-Augmented Generation Systems (Zurich)
About this job
Fixed-term contract (typically 3 months) · 80–100% · Zürich · Students (Bachelor, Master or PhD level) «Making sense of data is the winning competence in all industries.» We build D ONE with the belief that every company will be a data company. Since our beginnings, we help our clients to make sense of their data and generate value. In this internship you will work independently on a topic in the area of evaluating generative AI systems. You will turn fuzzy questions about quality into measurements that people trust, and make those measurements repeatable. Your tasks will include: • Reviewing evaluation methodology for retrieval-augmented and generative systems. • Designing and implementing automated evaluation logic in Python. • Building and curating test data sets, including gold standards and synthetic cases. • Analysing results statistically and turning them into clear, defensible conclusions. • Documenting your work and presenting it to a technical and a non-technical audience. What you bring: • Current enrolment in a Bachelor's, Master's or PhD programme in Computer Science, Data Science, Computational Linguistics, Engineering, Mathematics or a comparable field. • Interest in at least one industry domain and curiosity about how data creates value in it. • Down-to-earth and pragmatic results-oriented attitude. • Experience in working with large data sets and databases. • Knowledge in relevant technologies and programming languages (e.g. Python, SQL, Git, Spark, cloud platforms). • Excellent communication skills in English; German is a strong plus. Preferred qualifications: • Working understanding of large language models and retrieval-augmented generation, and of where such systems typically fail. • Familiarity with evaluation concepts such as faithfulness, groundedness, answer relevance, context precision and recall, and hallucination detection. • Exposure to evaluation tooling such as RAGAS, DeepEval, promptfoo, TruLens or LangSmith, or a willingness to build equivalent logic yourself. • Experience with LLM-as-a-judge techniques, and a healthy scepticism about their reliability. • Solid statistical foundations: sampling, significance testing, confidence intervals, inter-annotator agreement. • Practical skills in test-set construction, annotation guidelines and synthetic data generation. • Software engineering discipline: reusable Python packages, automated test suites, CI, reproducible experiment pipelines. • Ability to communicate quantitative results visually and concisely (e.g. pandas, plotly, matplotlib, dashboards). What we offer: • An uncomplicated and low noise environment that enables to focus on what matters most. • An international team of talented people from diverse backgrounds that have ambitions, always enjoy an additional challenge, and know how to have fun at work. • Direct mentoring by senior D ONE consultants, plus access to our D ONE Academy including knowledge sharing sessions, expert and working groups. • A clearly scoped topic that is yours to own, with the freedom to shape how you approach it. • Frequent offsite socialising events. • A realistic path to joining D ONE as a consultant after a successful internship. To be successful in this role you will need to: • Strive to understand your clients' needs. • Deliver outstanding results with high business value in a comprehensible form. • Have the intellectual agility required to get the job done in an easy-to-do-business-with way. • Value clever and creative team play. • Work independently and structure your own progress, while knowing when to pull in a senior colleague. Are you interested in joining a great team with a proven track record and to work on a topic that will constantly challenge you? We are happy to receive your application with reference: D ONE Value Creation AG Sihlfeldstrasse 58 8003 Zürich