• 2025: Led a team to develop the next version of our products (Foodscanner, FoodVision) and deployed pilots. • 2024: Developed a food waste monitoring system for airlines (FoodVision.io) using 3D cameras in airline caterer facilities with machine learning and computer vision. Technologies used were Python, Open3D, PyTorch for ML, TypeScript, RedwoodJS, React, GraphQL, Lambda functions, SQL for the web app. • 2023: Further developed both products (Foodscanner, Expensly) and onboarded customers in production. • 2022: Developed an expense management solution (Expensly.ch) complete with mobile app, web and backend. Technologies used were Flutter, Dart, Android, iOS for mobile, AWS Amplify, GraphQL, TypeScript, Lambda functions, NoSQL, REST, Stripe, OCR for web and backend. • 2021: Developed a patient monitoring system (Foodscanner) using 3D cameras in hospital kitchens with machine learning and computer vision. Technologies used were Python, Open3D, PyTorch for ML, TypeScript, RedwoodJS, React, GraphQL, Lambda functions, SQL for the web app.

• 2021: Led a team to develop a web app for Baloise to manage the whole claim management life cycle. Technologies used were C#.Net, React, Azure, REST. • 2020: Led a team to develop an end-to-end solution for the next generation industrial autonomous cleaning robot for Cleanfix including mobile app for management, API, cloud, HMI, autonomous charging and water refill. Technologies used were C#.Net, Python, React, Flutter, Dart, Azure, Docker, REST, Bluetooth. • 2019: Developed a mobile app for Baloise to manage claims directly at the customer. Technologies used were Flutter, Dart, Android, iOS. • 2018: Developed a mobile app for Cleanfix to see cleaning reports of an industrial autonomous cleaning robot including an API, web app, cloud deployment and Excel reporting. Technologies used were C#.Net, Flutter, Dart, Azure, REST.

• Built a frontend component library using React and TypeScript for internal tooling (https://github.com/open-ch/pyrene) • Focused on clean component structure and reusability

• Developed pipelines for food classification and audio segmentation using deep learning • Used Python, TensorFlow, Flask, Docker, and Vue.js • Focused on real-world healthcare data and performance

• Built object detection and segmentation pipelines for a surveying application • Used Mask R-CNN and 3D point cloud data to improve accuracy • Combined deep learning and 3D geometry for post-processing

• Maintained test environments for around 100 enterprise applications using Automic Workflow Automation, Mainframe Host and Unix shell • Automated testing for CRM, portfolio management, and mainframe systems using Tosca • Investigated and solved cross-system issues