Zurich, Zurich, Switzerland
Software Engineer specializing in Rust and C++ systems, currently working on security-critical custody platforms in regulated environments. My work focuses on designing, modernizing, and hardening backend systems where correctness, reliability, and security are non-negotiable. I have hands-on experience with PKCS#11 and HSM integration, Rust - C++ interoperability, and containerized production deployments. I enjoy tackling complex systems problems such as safely modernizing legacy codebases, reducing operational and security risk, and improving the performance and maintainability of long-lived services.
Worked on data processing pipelines and applied machine learning to better understand customer behavior at Swisscom Digital Lab. I enhanced data collection and ETL pipelines using Python and Apache Spark, transforming raw data from internal data lakes into structured datasets suitable for analysis and multi-task modeling. My work focused on feature engineering and evaluating the relevance of different signals across multiple prediction tasks. In collaboration with a Ph.D. researcher, I explored multi-task learning approaches and proposed a heuristic for grouping related tasks, improving prediction precision compared to single-task models. ● Implemented multi-task deep learning models (e.g., Cross-stitch Networks, Progressive Layer Extraction) and evaluated their performance on different tabular datasets. ● Proposed and implemented a multi-task learning method, improving precision by 3% compared to single-task models.
Worked on an end-to-end proof of concept for detecting known vulnerabilities (CVEs) in software packages using graph-based analysis. I designed and implemented a Java-based service leveraging Oracle Graph Server (PGX), taking ownership of the full delivery lifecycle from containerization to deployment on Kubernetes. This included integrating with cloud storage and data processing components to support scalable analysis. ● Identified an issue in Oracle’s HDFS connector during testing, proposing a fix to the relevant team.
Worked on improving deployment workflows for services running on an AWS EKS cluster, focusing on reducing duplication and simplifying operational complexity. I designed and implemented a general-purpose Jenkins pipeline to build, test, and deploy resources to Kubernetes, replacing multiple specialized pipelines. I also improved the team’s Slack-based deployment workflow by adapting an internal bot to support the new pipeline structure, making it easier for engineers to trigger deployments and manage resources independently. ● Reduced duplicate work and streamlined collaboration around service deployments by replacing approximately 10% of existing Jenkins jobs.