Tobias Krüger

Built AI-driven enrichment services using LLM APIs for large-scale hotel and destination content, including automated classification, generation, prompt design, output validation, and provider fallback handling. Designed Python ETL and normalization pipelines to consolidate heterogeneous supplier feeds into a consistent schema, with deduplication and record-matching rewritten in Rust (PyO3) for higher throughput. Owned REST API architecture for web clients and partner integrations, including versioning, authentication, rate limiting, and caching strategy. Delivered Java/Spring Boot services alongside the Python estate with Spring Security authentication, JPA persistence to PostgreSQL, and containerized deployment via the shared CI/CD pipeline. Developed Next.js/TypeScript frontends against Sanity structured content with improved fetching, caching, and rendering; drove architecture decisions, code review standards, testing, release process ownership, and engineer mentoring.

Built backend features for a post-purchase shipping and tracking SaaS, implementing multi-carrier API integration and normalizing tracking events across providers. Implemented an async Rust ingestion service (Axum, Tokio) for high-volume tracking data with retry, deduplication, and failure-recovery logic against unreliable carrier APIs. Connected merchant e-commerce systems to shop backends using REST APIs and webhooks to automate customer notification workflows.

Designed and built Spring Boot microservices for core insurance workflows, covering REST API design, Spring Security and OAuth2 authentication, Hibernate/JPA persistence against SQL databases, and JUnit test coverage. Developed and maintained backend services, internal APIs, and web applications in a regulated environment with strict security, data-protection, and code-review standards. Built integration layers between internal platforms and third-party services with data validation, error handling, and audit logging. Reduced operational workload by replacing recurring manual processes with Python tooling and scheduled jobs.