[{"data":1,"prerenderedAt":386},["ShallowReactive",2],{"jobs-en-bbv_engineer":3,"all-jobs-pages-en":68},{"id":4,"title":5,"audience":6,"body":7,"carouselItems":6,"companyName":52,"date":6,"description":53,"end_date":54,"eventDate":6,"eventDuration":6,"eventName":6,"extension":55,"highlightWord":6,"icon":6,"images":6,"label":6,"link":6,"logo":56,"mainTitle":6,"meta":57,"navigation":58,"path":59,"podcastId":6,"role":60,"seo":61,"series":6,"start_date":62,"stats":6,"stem":63,"subtitle":44,"tags":6,"techs":64,"titleHighlight":65,"videoId":6,"views":6,"watchTime":66,"__hash__":67},"jobs\u002Fjobs\u002Fen\u002Fbbv_engineer.md","From .NET Development to Deputy Mandate Lead",null,{"type":8,"value":9,"toc":43},"minimark",[10,14,19,22,26,29,33,36,40],[11,12,13],"p",{},"When I started at bbv I was initially working as a professional software developer in the Java space. In my very first project it quickly became clear that my expertise in .NET and C# was needed. I switched tech stacks at short notice, worked with a small agile team, and was able to deepen both my front-end experience and my C# skills. Within that project we successfully delivered the first development phase together.",[15,16,18],"h3",{"id":17},"taking-on-responsibility-mandate-lead-and-coordination","Taking on Responsibility: Mandate Lead and Coordination",[11,20,21],{},"After completing the initial development phase, I took on the role of deputy mandate lead from June 2023 for the support and ongoing development of the client mandate in the utilities billing domain (Blazor, .NET 7–9, Azure with Bicep Infrastructure-as-Code). My main focus was on communication with clients and stakeholders as well as coordinating the team. In this way I was able to demonstrate my leadership qualities and gain important project management experience.",[15,23,25],{"id":24},"data-science-community-and-co-leadership","Data Science Community and Co-Leadership",[11,27,28],{},"Alongside my project responsibilities I joined the Data Science community at bbv. Thanks to my enthusiasm for new technologies, and for artificial intelligence in particular, I quickly built trust within the group. A short time later I took on the co-leadership of this community. In this role I made a significant contribution to advancing AI competence across the firm.",[15,30,32],{"id":31},"building-an-ai-team-and-an-ai-division","Building an AI Team and an AI Division",[11,34,35],{},"Out of the Data Science community a dedicated AI team eventually emerged, which grew into its own AI division as internal interest expanded. A milestone was bringing my long-standing colleague and AI expert Joel Barmettler onto the team, which considerably raised our expertise and capability in AI development. The newly formed AI team built a technology platform that puts us in a position to design, develop, and deliver AI agents to clients quickly. On this foundation we have already realised numerous AI projects and prototypes, from intelligent chatbots to advanced analytics solutions.",[15,37,39],{"id":38},"promotion-to-senior-software-developer-in-2024","Promotion to Senior Software Developer in 2024",[11,41,42],{},"In the course of my professional development at bbv I was promoted to Senior Software Developer in 2024. This recognition underlines my several years of experience, my technical expertise, and my ability to steer projects successfully from both a technical and an organisational perspective. I also had the opportunity to represent bbv twice at the .NET Conference in Porto and London, broadening my technical perspective beyond my own mandate.",{"title":44,"searchDepth":45,"depth":45,"links":46},"",2,[47,49,50,51],{"id":17,"depth":48,"text":18},3,{"id":24,"depth":48,"text":25},{"id":31,"depth":48,"text":32},{"id":38,"depth":48,"text":39},"bbv Software Services AG","Joined bbv as a Java engineer, switched quickly to the .NET stack. Multi-year client mandate in the utilities billing domain (Blazor, .NET, Azure), took on deputy mandate lead in 2023, promoted to Senior in 2024. Parallel co-leadership of the Data Science community and co-founding of the bbv AI team.","2024-12-31","md","\u002Fimages\u002Fjobs\u002Fbbv.png",{},true,"\u002Fjobs\u002Fen\u002Fbbv_engineer","Software Engineer (.NET) → Senior, Deputy Mandate Lead",{"title":5,"description":53},"2022-08-01","jobs\u002Fen\u002Fbbv_engineer",".NET, Blazor, Azure, Bicep, Cypress","bbv","56mins","25Kl-ynrpuFT9WMlTNYw2srpCvwtn113WFXP8fb26Xs",[69,120,149,226,276,331],{"id":70,"title":71,"audience":6,"body":72,"carouselItems":6,"companyName":52,"date":6,"description":111,"end_date":6,"eventDate":6,"eventDuration":6,"eventName":6,"extension":55,"highlightWord":6,"icon":6,"images":6,"label":6,"link":6,"logo":56,"mainTitle":6,"meta":112,"navigation":58,"path":113,"podcastId":6,"role":114,"seo":115,"series":6,"start_date":116,"stats":6,"stem":117,"subtitle":44,"tags":6,"techs":118,"titleHighlight":65,"videoId":6,"views":6,"watchTime":66,"__hash__":119},"jobs\u002Fjobs\u002Fen\u002Fbbv_ai.md","Senior AI Engineer and AI Consultant",{"type":8,"value":73,"toc":105},[74,77,81,84,88,91,95,98,102],[11,75,76],{},"Since the beginning of 2025 I have been fully dedicated to the topic of artificial intelligence at bbv. My main responsibility is to design forward-looking AI solutions and bring them into productive use at client sites. In this role I support companies in their digital transformation, optimise business processes, and create real value through targeted AI and data science solutions.",[15,78,80],{"id":79},"co-engineer-of-the-swiss-ai-hub-platform","Co-Engineer of the Swiss AI Hub Platform",[11,82,83],{},"At the centre of my work is the Swiss AI Hub, a transparency- and sovereignty-driven open-source AI platform. As co-engineer and solution architect I am responsible for core components including the agent runtime, the RAG pipeline, and the evaluation and tracing infrastructure. Technologies in use include Python, FastAPI, FastMCP, Milvus, Neo4j, LiteLLM, LlamaIndex, Langfuse, Keycloak, NATS, PostgreSQL, OpenTelemetry, and OpenInference, operated on Docker and OpenStack with Ansible-based automation.",[15,85,87],{"id":86},"technical-lead-on-client-projects","Technical Lead on Client Projects",[11,89,90],{},"I accompany client projects from use-case identification in workshops through proof of concept to MVP. This includes architecture and security concepts as well as the selection of suitable models and infrastructure components. I also act as a technical expert in pre-sales and have contributed to two national tenders for AI platforms.",[15,92,94],{"id":93},"knowledge-transfer-webinars-and-talks","Knowledge Transfer: Webinars and Talks",[11,96,97],{},"Since late 2023 I have been delivering regular webinars, workshops, and talks on LLMs, AI Agents, the Model Context Protocol (MCP), and AI infrastructure. I have now accumulated more than 30 appearances, including a talk at the Azure Cosmos DB Online Conference 2025 with around 2,000 participants, contributions to the AI practice circle of Next Industries, and sessions at AI bootcamps at FHNW. In parallel I share my knowledge in client projects and internal training sessions, so that both clients and colleagues benefit from the latest trends and developments.",[15,99,101],{"id":100},"ai-in-software-development","AI in Software Development",[11,103,104],{},"Alongside working on AI products, I make strong use of AI-assisted development myself, particularly with Claude Code (skills, agents, hooks). This brings the experience of being an end user directly back into the architecture and tooling of the platform.",{"title":44,"searchDepth":45,"depth":45,"links":106},[107,108,109,110],{"id":79,"depth":48,"text":80},{"id":86,"depth":48,"text":87},{"id":93,"depth":48,"text":94},{"id":100,"depth":48,"text":101},"Fully dedicated to AI since early 2025: co-engineer of the Swiss AI Hub platform (agent runtime, RAG pipeline, evaluation, tracing), technical lead on client projects from workshop to MVP, pre-sales as technical expert, and more than 30 talks on LLMs, AI Agents, and AI infrastructure.",{},"\u002Fjobs\u002Fen\u002Fbbv_ai","Senior AI Engineer & AI Consultant",{"title":71,"description":111},"2025-01-01","jobs\u002Fen\u002Fbbv_ai","Python, FastAPI, MCP, LlamaIndex, Milvus, Langfuse","SWfmxBUfrTsj73PuQ_UmnVvcfFI3uOTIkgLcPlwUhqM",{"id":4,"title":5,"audience":6,"body":121,"carouselItems":6,"companyName":52,"date":6,"description":53,"end_date":54,"eventDate":6,"eventDuration":6,"eventName":6,"extension":55,"highlightWord":6,"icon":6,"images":6,"label":6,"link":6,"logo":56,"mainTitle":6,"meta":147,"navigation":58,"path":59,"podcastId":6,"role":60,"seo":148,"series":6,"start_date":62,"stats":6,"stem":63,"subtitle":44,"tags":6,"techs":64,"titleHighlight":65,"videoId":6,"views":6,"watchTime":66,"__hash__":67},{"type":8,"value":122,"toc":141},[123,125,127,129,131,133,135,137,139],[11,124,13],{},[15,126,18],{"id":17},[11,128,21],{},[15,130,25],{"id":24},[11,132,28],{},[15,134,32],{"id":31},[11,136,35],{},[15,138,39],{"id":38},[11,140,42],{},{"title":44,"searchDepth":45,"depth":45,"links":142},[143,144,145,146],{"id":17,"depth":48,"text":18},{"id":24,"depth":48,"text":25},{"id":31,"depth":48,"text":32},{"id":38,"depth":48,"text":39},{},{"title":5,"description":53},{"id":150,"title":151,"audience":6,"body":152,"carouselItems":6,"companyName":213,"date":6,"description":214,"end_date":215,"eventDate":6,"eventDuration":6,"eventName":6,"extension":55,"highlightWord":6,"icon":6,"images":6,"label":6,"link":6,"logo":216,"mainTitle":6,"meta":217,"navigation":58,"path":218,"podcastId":6,"role":219,"seo":220,"series":6,"start_date":221,"stats":6,"stem":222,"subtitle":44,"tags":6,"techs":223,"titleHighlight":224,"videoId":6,"views":6,"watchTime":66,"__hash__":225},"jobs\u002Fjobs\u002Fen\u002Fpolygon.md","Founding My Own Software Startup",{"type":8,"value":153,"toc":205},[154,158,161,165,168,172,175,178,182,185,189,192,196,199,202],[15,155,157],{"id":156},"the-beginning-of-an-idea","The Beginning of an Idea",[11,159,160],{},"During my studies at the Universität Zürich the idea emerged, together with two of my fellow students, to start our own company. Everything began at a table on the second floor of the Institute of Computer Science on Binzmühlestrasse. Without any fanfare, but with a dose of courage and a lot of enthusiasm, we quickly sketched out a first concept, written with a digital stylus in an unnamed OneNote file. Between lecture notes and shopping lists, our simple but ambitious business model took shape.",[15,162,164],{"id":163},"a-generation-with-potential","A Generation with Potential",[11,166,167],{},"We recognised that our generation of students, growing up with the latest technologies, not only expected modern working conditions such as flexible hours, flat hierarchies, and varied tasks, but also possessed enormous potential. We wanted to harness that potential and at the same time create a working environment where students and young talent could feel at home and flourish.",[15,169,171],{"id":170},"an-unconventional-start","An Unconventional Start",[11,173,174],{},"So we decided to launch a software development company that was not oriented around the rigid patterns of traditional businesses but was built from the outset around the needs and interests of students. Our goal was to develop innovative software solutions using cutting-edge technologies while simultaneously creating a unique working environment.",[11,176,177],{},"Our start was anything but easy. We had no extensive professional experience, so we had to learn everything: How do we communicate as a team? How do we structure our working hours? How do we write proposals for clients who often did not know exactly what they needed? How do we set our hourly rate? Every question, however small, was answered through research, trial and error, and intensive brainstorming. In the process we frequently developed or adapted the tools and aids we needed ourselves. These challenges forced us to find innovative and timely solutions that were often unique.",[15,179,181],{"id":180},"challenges-and-growth","Challenges and Growth",[11,183,184],{},"As founders we grew into our new roles. We learned to conduct job interviews, run meetings, and present to clients who had decades more experience than we did. The harder decisions, such as letting people go, were also part of it. Despite all the challenges we experienced dynamic development: our small idea became a GmbH with a modern office right next to the computer science campus in Oerlikon. Over time we moved twice into larger and more modern offices, always staying in the neighbourhood of the Universität Zürich in Oerlikon, to remain close to our most valuable asset: the students. Within two years our team grew from three to 18 people, and we won numerous clients while simultaneously running more than ten projects.",[15,186,188],{"id":187},"continuous-improvement","Continuous Improvement",[11,190,191],{},"Despite our success, Polygon Software kept moving forward. Our focus was on optimising processes, improving methods, and automating repetitive tasks. We remained true to our founding idea: our employees, mainly students and alumni of the Universität Zürich, were at the centre of everything. Flexible working conditions, such as unlimited unpaid leave or working from home, were as natural to us as the ability to reduce hours during exam periods. Many of our staff were friends from university or from our personal lives who saw themselves as part of a startup in which they played a meaningful role and could find personal fulfilment.",[15,193,195],{"id":194},"a-personal-turning-point","A Personal Turning Point",[11,197,198],{},"Over time we faced more and more challenges in which management and business concerns increasingly dominated, while software development and technology, my actual passion, receded further into the background. As founders we had consciously built the company according to our own vision rather than rigid conventions. But that is precisely what led me to reflect: were we still on the right path?",[11,200,201],{},"I noticed increasingly that I saw myself less as a managing director and more as someone who wanted to focus on technical work. After finishing my studies I worked full-time for the company for a year, often well beyond 100% if one can put it that way. At some point it was time to step back and gain some perspective. I decided on a three-week trip to find clarity.",[11,203,204],{},"During that trip it became clear to me that I needed new perspectives and wanted to see something different. After much deliberation I made the extremely difficult decision to leave the company. It was not just the company I had helped build; it was also my close friends, whom I had to leave behind in a certain sense. That step was one of the hardest of my life.",{"title":44,"searchDepth":45,"depth":45,"links":206},[207,208,209,210,211,212],{"id":156,"depth":48,"text":157},{"id":163,"depth":48,"text":164},{"id":170,"depth":48,"text":171},{"id":180,"depth":48,"text":181},{"id":187,"depth":48,"text":188},{"id":194,"depth":48,"text":195},"PolygonSoftware","Co-founder of a software startup launched in 2019 during my studies at the Universität Zürich. Focus on innovative software solutions and a student-friendly work environment. After more than three years and significant growth I decided to leave the company and explore new directions.","2022-05-31","\u002Fimages\u002Fjobs\u002FPolygonSoftware.png",{},"\u002Fjobs\u002Fen\u002Fpolygon","Co-founder & Tech Lead",{"title":151,"description":214},"2019-03-01","jobs\u002Fen\u002Fpolygon","Vue.js, AWS, Typescript","Software","XgJOrbGkLT2cYx07TcNvZF3D8KW6vOXmsxqY2x6sbzM",{"id":227,"title":228,"audience":6,"body":229,"carouselItems":6,"companyName":263,"date":6,"description":264,"end_date":265,"eventDate":6,"eventDuration":6,"eventName":6,"extension":55,"highlightWord":6,"icon":6,"images":6,"label":6,"link":6,"logo":266,"mainTitle":6,"meta":267,"navigation":58,"path":268,"podcastId":6,"role":269,"seo":270,"series":6,"start_date":271,"stats":6,"stem":272,"subtitle":44,"tags":6,"techs":273,"titleHighlight":274,"videoId":6,"views":6,"watchTime":6,"__hash__":275},"jobs\u002Fjobs\u002Fen\u002Ftracktics_box.md","TRACKTICS Desktop App for Data Synchronisation with the Trainer Box",{"type":8,"value":230,"toc":258},[231,234,238,241,245,248,252,255],[11,232,233],{},"TRACKTICS is known for its innovative football performance trackers, offered both as individual devices and in a box containing multiple units. This box makes it possible to collect training data in one place. For that data to be used for analysis, however, it needs to be transferred from the box to the cloud. An existing desktop application provided the foundation for this, but its functionality and reliability had room for improvement.",[15,235,237],{"id":236},"the-challenge-wi-fi-connections-and-error-management","The Challenge: Wi-Fi Connections and Error Management",[11,239,240],{},"My goal was to extend the desktop application and turn it into a robust and user-friendly tool. The focus was on error management, since Wi-Fi connections were a recurring source of complications. These challenges affected not only the technical side but also the user experience. Together with a designer I developed a new UI\u002FUX concept that guides the user step by step through the data transfer process, from establishing the connection to a successful upload to the cloud.",[15,242,244],{"id":243},"technical-solution-and-collaboration","Technical Solution and Collaboration",[11,246,247],{},"Developing the desktop application required a deep understanding of the box's firmware. Through close collaboration with the firmware team I was able to ensure that the application communicates smoothly with the hardware. I also placed particular emphasis on cross-platform usability. The application was optimised for Windows and macOS and equipped with an automatic update function. This ensures that users always have the latest version of the software without having to search for updates manually.",[15,249,251],{"id":250},"results-and-outlook","Results and Outlook",[11,253,254],{},"My work enabled TRACKTICS to deliver an improved version of the desktop application that is not only more stable but also offers a better user experience. The automatic update function ensures that customers always benefit from the latest improvements. I was also able to contribute valuable insights and ideas for the ongoing development of the software.",[11,256,257],{},"This enhancement shows how important a combination of technical expertise and user focus is for making innovative products even better. It is a further step in positioning TRACKTICS as a pioneer in football performance analysis.",{"title":44,"searchDepth":45,"depth":45,"links":259},[260,261,262],{"id":236,"depth":48,"text":237},{"id":243,"depth":48,"text":244},{"id":250,"depth":48,"text":251},"TRACKTICS GmbH","Development and optimisation of the TRACKTICS desktop app with a focus on Wi-Fi error management, cross-platform deployment, and automatic updates for a smooth user experience.","2020-06-31","\u002Fimages\u002Fjobs\u002FTracktics.png",{},"\u002Fjobs\u002Fen\u002Ftracktics_box","Working Student",{"title":228,"description":264},"2017-11-01","jobs\u002Fen\u002Ftracktics_box","JavaScript, Vue.js, Electron, Zeplin, TestRail","TRACKTICS","yCXOvIzFuBqfRerDzlb75YexLrBr9k-dkdIvn7Iow58",{"id":277,"title":278,"audience":6,"body":279,"carouselItems":6,"companyName":263,"date":6,"description":321,"end_date":322,"eventDate":6,"eventDuration":6,"eventName":6,"extension":55,"highlightWord":6,"icon":6,"images":6,"label":6,"link":6,"logo":266,"mainTitle":6,"meta":323,"navigation":58,"path":324,"podcastId":6,"role":325,"seo":326,"series":6,"start_date":327,"stats":6,"stem":328,"subtitle":44,"tags":6,"techs":329,"titleHighlight":274,"videoId":6,"views":6,"watchTime":6,"__hash__":330},"jobs\u002Fjobs\u002Fen\u002Ftracktics_praktikum.md","Internship: Detecting Jogging Sessions in Football Tracker Data",{"type":8,"value":280,"toc":315},[281,284,288,291,295,298,301,305,308,312],[11,282,283],{},"TRACKTICS tracking devices are true all-rounders.\nThey capture GPS data, record movement patterns, and are primarily intended for analysing football training sessions. The flexibility of their users, however, brought an unexpected challenge: the devices were used not only on the pitch but also for jogging. This created a problem for data analysis, because jogging and football activities differ fundamentally, both in their dynamics and in the focus of analysis. A solution was needed to distinguish these session types clearly.",[15,285,287],{"id":286},"a-look-behind-the-scenes","A Look Behind the Scenes",[11,289,290],{},"When I took on this task it was clear that the key lay in the data processing pipeline. In a UNIX-based system I first had to familiarise myself with the existing process, understand it in depth, and find the right place to intervene. I noticed that the movement patterns offered a valuable source of information. Direction changes, speed variations, and the frequency of such events all reveal whether someone is chasing quick plays on the pitch or accumulating kilometres on a running track.",[15,292,294],{"id":293},"the-birth-of-the-algorithm","The Birth of the Algorithm",[11,296,297],{},"Armed with this knowledge, I developed an algorithm that analyses the frequency of direction and speed changes and derives the session type from them. Theory alone is not enough, though. To make sure the algorithm held up in practice, I applied data science methods to test and validate the reliability of its decisions.",[11,299,300],{},"The result was impressive: the algorithm proved itself not only in the test environment but was also integrated seamlessly into the production pipeline. It now reliably identifies whether a session was a jogging run or a football training, laying the groundwork for specific and targeted data analysis.",[15,302,304],{"id":303},"what-this-means-for-users","What This Means for Users",[11,306,307],{},"With the new solution the TRACKTICS platform becomes even more valuable to its users.\nData can now be evaluated in a way that is tailored to the session type. Football training sessions receive a dynamic analysis focused on game situations, while jogging data puts endurance performance front and centre. These results are presented clearly and attractively in the user interface, improving the user experience and creating genuine added value.",[15,309,311],{"id":310},"a-step-forward","A Step Forward",[11,313,314],{},"This development shows how important it is to respond flexibly to user needs and to solve technical challenges creatively. The algorithm is more than just a piece of code: it is a tool that brings order to data chaos and takes the TRACKTICS platform to a new level.",{"title":44,"searchDepth":45,"depth":45,"links":316},[317,318,319,320],{"id":286,"depth":48,"text":287},{"id":293,"depth":48,"text":294},{"id":303,"depth":48,"text":304},{"id":310,"depth":48,"text":311},"Algorithm developer for TRACKTICS tracking devices. Development of an algorithm for the automatic detection of jogging and football activities by analysing movement patterns. Integration into UNIX-based data pipelines to optimise the user experience and enable data-driven insights.","2017-10-31",{},"\u002Fjobs\u002Fen\u002Ftracktics_praktikum","Intern",{"title":278,"description":321},"2017-07-01","jobs\u002Fen\u002Ftracktics_praktikum","Python, Jupyter, Docker","7GXq8OThJxSuqRGlDPuERLne1t5IFRQ7iUjFaL2Igyw",{"id":332,"title":333,"audience":6,"body":334,"carouselItems":6,"companyName":373,"date":6,"description":374,"end_date":375,"eventDate":6,"eventDuration":6,"eventName":6,"extension":55,"highlightWord":6,"icon":6,"images":6,"label":6,"link":6,"logo":376,"mainTitle":6,"meta":377,"navigation":58,"path":378,"podcastId":6,"role":379,"seo":380,"series":6,"start_date":381,"stats":6,"stem":382,"subtitle":44,"tags":6,"techs":383,"titleHighlight":384,"videoId":6,"views":6,"watchTime":6,"__hash__":385},"jobs\u002Fjobs\u002Fen\u002Fuzh.md","Research Engineer in the Computation and Economics Research Group",{"type":8,"value":335,"toc":367},[336,339,343,346,350,353,357,360,364],[11,337,338],{},"After completing my master's degree I worked as a research assistant in the Computation and Economics Research Group of Prof. Dr. Sven Seuken at the Universität Zürich from January 2021 to April 2022. The core period was May to October 2021, when I worked as a full-time research engineer with the group after submitting my master's thesis.",[15,340,342],{"id":341},"research-on-neural-networks-and-bayesian-optimisation","Research on Neural Networks and Bayesian Optimisation",[11,344,345],{},"My substantive focus was on applying neural networks to optimisation problems, in particular in the context of Bayesian optimisation with model-based uncertainty estimates. I was able to build directly on my master's thesis \"Bayesian Optimization with Uncertainty Bounds for Neural Networks\" and bring my results into a broader research context. I am named in the acknowledgements of the publication \"NOMU: Neural Optimization-based Model Uncertainty\".",[15,347,349],{"id":348},"codebase-refactoring-from-functional-to-object-oriented","Codebase Refactoring from Functional to Object-Oriented",[11,351,352],{},"A significant contribution was carefully migrating the historically grown research codebase from a strongly functional style towards an object-oriented architecture. Through the strategy pattern, clear interfaces, and cleanly separated components, the code became considerably more maintainable, testable, and extensible for ongoing research, without breaking established experiments.",[15,354,356],{"id":355},"reproducible-experiment-pipeline-on-slurm","Reproducible Experiment Pipeline on Slurm",[11,358,359],{},"In parallel I built a reproducible experiment pipeline on a Slurm CPU cluster. This allowed hyperparameter sweeps, comparison runs, and ablation studies to be systematically queued, evaluated in an automated fashion, and stored with clean versioning. The primary tools were Python, scipy, TensorFlow, Keras, and Git.",[15,361,363],{"id":362},"transition-to-industry","Transition to Industry",[11,365,366],{},"My time at UZH shaped my understanding of well-structured research software as well as the use of neural networks in real-world optimisation problems. I subsequently decided to deepen this experience in industry.",{"title":44,"searchDepth":45,"depth":45,"links":368},[369,370,371,372],{"id":341,"depth":48,"text":342},{"id":348,"depth":48,"text":349},{"id":355,"depth":48,"text":356},{"id":362,"depth":48,"text":363},"Universität Zürich","Research in the Computation and Economics Research Group of Prof. Dr. Sven Seuken on neural networks and Bayesian optimisation. Refactoring the research codebase from a functional to an object-oriented style and building a reproducible experiment pipeline on a Slurm CPU cluster. Named in the acknowledgements of the publication «NOMU: Neural Optimization-based Model Uncertainty».","2022-04-30","\u002Fimages\u002Fjobs\u002Fuzh.png",{},"\u002Fjobs\u002Fen\u002Fuzh","Research Assistant (Research Engineer)",{"title":333,"description":374},"2021-01-01","jobs\u002Fen\u002Fuzh","Python, TensorFlow, Keras, scipy, Slurm","Zürich","6mxwrdP-QAIoHBU1f_lRrOxmGYt0VPRT1gDfd6P7W-s",1779977587617]