Rasmus Risse

Replaces the previously manual RFQ pricing in energy trading, where traders had to research and price each incoming request by hand under time pressure. Concept and implementation of an automated pricing engine with seamless integration of the existing trading systems via gRPC and REST APIs built to the OpenAPI specification. Delivery of a high-performance messaging layer on Aeron and Akka .NET for real-time processing and resilient communication between trading components. Operation as microservices on Kubernetes with PostgreSQL as the persistent data store, continuously rolled out via Azure DevOps, Helm, and Terraform. Frees up trading capacity, reduces errors from manual entries, and noticeably shortens response times to incoming requests.

Profit optimization of an existing gas storage platform for RWE Supply & Trading, whose outdated stack had become a bottleneck for rolling out new trading strategies into production. Full modernization of the backend to the current .NET version with Entity Framework, consolidation of data models, and redesign of the Angular user interfaces to match corporate identity. Build-out of a horizontally scalable microservices infrastructure on Azure (Docker, Kubernetes, Terraform) including Azure DevOps pipelines and comprehensive automated tests. Securing all internal and external interfaces via Microsoft Entra ID, REST/OpenAPI APIs, and messaging through Amazon SQS. Creates the foundation for future trading strategies and features, and noticeably raises availability in live gas trading operations.

Modernization and continuous development of a large-scale portfolio management platform for long-term gas trading at RWE Supply & Trading, whose recurring instabilities disrupted production operations and slowed down trading activity. Refactoring of critical backend components in C#/.NET, hardening of the data layers connecting MSSQL, MongoDB, and PostgreSQL, and performance optimization of the Angular user interfaces with AG Grid and NgRx. Integration of additional trading modules via REST/OpenAPI APIs and messaging through Amazon SQS, maintenance of the containerized microservices on Kubernetes with Azure DevOps and Terraform. Stabilizes live trading operations, shortens release cycles, and noticeably improves the pace at which portfolio managers can work with large trading books.

Internal IDENTIC initiative to replace recurring manual processes with AI-assisted workflows. The starting point was a set of routine tasks tying up significant capacity, while isolated LLM tools only flowed into daily work via copy-paste. Design and build of an n8n-based automation platform with TypeScript and Node.js extensions, integrating Anthropic Claude and OpenAI through standardized APIs. Implementation of RAG pipelines (Retrieval-Augmented Generation) and MCP integration (Model Context Protocol) as reusable building blocks for context-aware, language-driven automation. Design of a scalable, containerized architecture (Docker, REST APIs) that allows fast model or tool swaps without deep refactoring. Noticeably reduces process times in the automated areas, frees teams from routine work, and establishes a reusable pattern for new AI use cases.

Greenfield build of the e-commerce platform for Geschenkzeitung.de, a niche shop offering original historic newspapers as gift items. The starting point was a set of manual, labor-intensive processes across catalog management, order handling, and product search, which were barely sustainable in headcount terms given a range of more than two million individual items. Software architecture design and implementation of a Progressive Web App with Angular, TypeScript, and NgRx, together with a Node.js/Express backend on MongoDB, complemented by REST integrations to payment providers and product systems. Delivery of performant search, filter, and checkout flows, offline-capable PWA UX, and a mobile-first interface. Result: a shop that scales cleanly to 2M+ items, a fast path from MVP into production, and a noticeable reduction of manual effort through digitization of operational processes.

Greenfield build of a custom e-commerce platform for a B2C food startup, where shop, inventory management, and food-specific production processes had to interlock tightly. The starting point was that off-the-shelf shop systems could not reflect the individual processes across production, shelf life, and fulfillment. Implementation of an Angular frontend with NgRx, RxJS, and TypeScript, together with a Node.js/Express backend on MongoDB as microservices, complemented by a custom inventory management system and REST integrations between shop, warehouse, and production. Containerization via Docker, deployment on Google Cloud Platform with Cloud Run and Cloud Storage, CI/CD through Azure DevOps. Result: a fully integrated end-to-end stack from product catalog to delivery, a fast path from concept into production, and a cloud infrastructure that scales automatically under variable load.

On-premise software for the analysis of microbiological urine diagnostics according to EUCAST standards as part of the ProDocMed project UroStatistix. The background was a regulatory requirement for urology practices and laboratories to document antibiogram resistances completely and traceably; until then this had been almost entirely manual and barely auditable. Responsibility for requirements analysis with the domain stakeholders, functional design of the statistical evaluation logic, and build-out of the backend with Node.js, Express, and MongoDB. Implementation of the Angular frontend for data entry, reports, and audit history, packaged as a local Windows installation via Inno Setup; a complementary online component (e.g., for order handling) runs in Docker. Automates EUCAST-compliant report generation, saves documentation time in practices and laboratories, and creates an auditable data foundation.

Development of a custom ERP solution for a client who considered standard systems such as SAP or Odoo too generic and too expensive for their specific processes. The starting point was a landscape of disconnected point solutions and manual routines across HR, inventory, projects, and fleet, with neither end-to-end coverage nor reliable reporting. Application architecture design and implementation of a C#/.NET desktop application with a WPF/XAML UI and MSSQL data layer, structured into four core modules (HR, inventory, project management, fleet management). System design and documentation with UML and BPMN. Responsibility as lead developer with a small development team, complemented by on-site rollout and end-user training in China. Result: a precisely tailored, significantly more affordable alternative to standard ERP systems, consolidating previously scattered processes into one central platform.

Development of a desktop client for securely reading, verifying, and managing data stored on ID smartcards, commissioned by a government client. The starting point was a set of heterogeneous administrative workflows, in which card access varied by device and location and was only partially secure. Implementation of card communication via APDU commands per ISO 7816 in C++ and Java, cryptographic verification of card data against ICAO security standards through PKI signature validation, and delivery of a user-friendly interface for case workers. Integration into the existing government infrastructure, including on-site installation and active on-site support during rollout. Result: a unified, ICAO-compliant client software that turned previously manual and inconsistent smartcard workflows into a secure, auditable process.

Development of software for authenticating passports and national ID documents in accordance with the ICAO standard, aimed at operational use in border control and security-sensitive environments. The starting point was a set of manual, error-prone verification processes, in which forgeries were hard to detect and slow to confirm. Implementation of the cryptographic verification chain in C++ and Java (OpenSSL, Bouncy Castle) for PKI-based signature validation, of RFID/NFC communication with the document chip in accordance with ISO 14443 (BAC/PACE), and of a JavaFX desktop client for manual inspection by security personnel, including integration into existing security and administrative infrastructures. Continuous responsibility over four years across crypto logic, chip communication, UI, and rollout. Result: productive deployment at several international airports, noticeably more reliable forgery detection, and an invited presentation of the technology at the EU conference at the Joint Research Centre in Ispra on document forgery prevention.

Intraoperative localization of brain tumors poses a clinical challenge: established modalities such as MRI or CT are expensive and often not readily available during surgery, so clinicians need a low-cost, non-invasive imaging approach to support neurosurgeons. As part of a bachelor's/master's thesis at the medical research chair of Ruhr University Bochum, a C++ research software was designed and implemented to analyze contrast-enhanced ultrasound (CEUS) videos and approximate a parametric perfusion model to the temporal dynamics of the contrast agent; model parameters are optimized via evolutionary algorithms, video processing and image analysis build on OpenCV, and algorithmic prototyping was performed in MATLAB. The pipeline trains detection models to localize suspected tumor regions, with particular attention to the temporal inflow behavior of the contrast agent. The work combines medical image processing, evolutionary optimization, and classical numerical algorithms, contributing to research on non-invasive brain tumor diagnostics and, prospectively, to intraoperative decision support.