Data Scientist (Python)

About us

SMARTSEER—now part of GIATA—builds a personalization platform for ecommerce that turns behavioral signals into real-time decisions across user experiences such as search, recommendations, and churn prevention. Our stack blends data pipelines, predictive user profiling, and APIs that deliver cutting-edge AI solutions at lightning speeds. Proven in travel and designed to generalize, we give engineers real ownership: shaping products, solving hard problems at scale, and seeing their work land in production fast.

Why this role

We have a rising demand on AI-related use cases and client requests. You’ll face machine learning use cases in an end-to-end context using Python and Microsoft Cloud infrastructure / Azure ML. Concept, feature engineering, model pipeline, evaluation, deployment, maintenance, automation, A/B tests - everything is in your hands. In addition you use your knowledge and skills to support colleagues with your insights on the data.

What you'll do

• Build end-to-end AI model pipelines: data export, preprocessing, transformation, training and evaluation.
• Perform MLOps: Deployment, Availability & Scalability of endpoints, Automation of pipelines.
• Raise data quality: Identify (and handle) issues in various data sources and structures and request additional ones.
• Handle data & profiles: work with user/trip summaries and predictions.
• Get insights: Apply NLP and CV approaches (tracking, product data, reviews, images).
• Improve our data platform: Running A/B tests, improving ML evaluation and monitoring capabilities, merging model output with business requirements (steering).
• Continuous improvement: Research for latest algorithms and approaches to improve existing use cases or tackle new ones.
• Strengthen the team: Be part of the Data Science team and collaborate with colleagues and clients.

Working with LLMs (pragmatic & secure)

We treat LLMs as force multipliers, not autopilots. Use them to draft design options/ADRs, enumerate edge cases, scaffold tests/fixtures, generate OpenAPI clients/boilerplate, and summarize PRs—then review, benchmark, and monitor before shipping. Privacy & safety first: no secrets or client data in prompts; stick to approved tools/policies.

What makes you a great fit (must-haves)

• Strong knowledge of Machine Learning concepts and state-of-the-art Python libraries.
• Knowledge in Cloud-based environments and their capabilities, preferred Azure ML.
• Very deep knowledge in ML / data manipulation (pandas, numpy, scikit-learn, xgboost).
• Applied experience in GPU-accelerated / parallel processing (pytorch, transformers, dask, polars, rapids).
• Experience in processing GB-scale datasets.
• Evaluate and visualise results using matplotlib, seaborn and SHAP.
• Exploratory mindset: Strong motivation for researching new approaches and creating concepts.
• Clear, kind communication in a small, collaborative, English-speaking team.

Great signals (nice to have)

• Interest about latest improvements in ML-related software ecosystem and hardware and experience with Huggingface and Kaggle.
• Knowledge in databases like MongoDB, elasticsearch and Cassandra.
• Kubernetes / AKS.
• Configuration of endpoints and VMs / MLOps.
• Experience with data-rich products (analytics, recommendations, personalization) or model serving.

Our stack (you'll touch some of this)

• Backend: Java microservices; Python for AI/ML models; targeted TypeScript/Node where it fits; OpenAPI-first contracts with generated clients.
• Frontend: React, TypeScript, modern CSS, Storybook.
• Platform: Java & Python data/ML services; analytics & profiling pipelines.
• DevOps/Security: Azure DevOps (Ansible/Helm), Docker, Kubernetes on Azure, SonarCloud, Snyk; SSO via Google/Microsoft Entra ID.
• Quality & Observability: Jest/RTL for web, contract/integration tests for services, ESLint/Prettier, Sentry, metrics/tracing.

Work style & location

• Hybrid, Frankfurt am Main — typically ~2–3 days/week together in the office.
• English-only in the office. No German required.
• No on-call / no out-of-hours support expectations.

What we offer

• Multicultural startup culture within an established travel-tech group (GIATA).
• Flat hierarchies, high trust, and real autonomy.
• Scope to shape architecture, tooling, and team practices.
• 30 days vacation.
• Modern tools (IntelliJ/WebStorm, sensible CI/CD).