Consulting and development for various clients mainly on cloud development, machine learning and software engineering

At Oxolo, I played a key role in shaping the technical foundation that enabled the company to secure a €13M Series A investment, as highlighted by TechCrunch: https://techcrunch.com/2023/10/24/oxolo-bags-e13m-for-gen-ai-driven-video-platform-which-can-optimise-engagement-on-the-fly/ . My work centered on driving innovation, scaling infrastructure, and building AI solutions that directly impacted Oxolo’s ability to deliver cutting-edge video creation tools.*Architected and scaled Oxolo's core AI platform: Designed and developed the cloud platform, APIs, and AI services, transforming an early-stage product into a robust, reliable, and maintainable solution that powers Oxolo’s generative video platform.*Established an MLOps-driven culture: Defined and implemented best practices for MLOps and DevOps, enabling seamless collaboration between research and engineering teams to accelerate the transition from research to production.*Transformed vime.ai into oxolo.com: Led the migration and refactoring of Oxolo’s initial platform, ensuring a smooth transition to its next-generation product offering.*Delivered scalable deployment solutions: Built and managed the AI deployment infrastructure using Cortex.dev, and introduced developer tooling like DVC and Poetry to improve workflow efficiency and maintainability.*Pioneered the video creation engine: Conceptualized and engineered key components of Oxolo’s video creation infrastructure, including advanced job scheduling systems, which became critical to the platform's functionality.*Drove innovation for product growth: My contributions directly supported Oxolo’s ability to attract investors and scale, playing a pivotal role in the company’s strategic trajectory and technological evolution.

* Developed the cloud platform, API and AI services behind oxolo and refactored the previous software design into a reliable and maintainable product.* Established an efficient MLOps culture to build and ship AI products. * Established best practices for MLOps and DevOps to transform research into products.* Built the cloud platform, API and AI services behind vime.ai from concept to product and impacted the choice of tech to ship faster.

* Conversational AI Platform for ecommerce based in Copenhagen (working 100% remotely from Germany)* Responsible for all AI product features and definition of roadmap of all the AI capabilities in the company. Furthermore, deciding on architectural decisions of the solution providing a holistic and scalable approach and mentoring junior/mid-developers. Leading a small team of analysts and engineers.* Tech: Python, FastAPI, PyTorch, Terraform, AWS, Azure, Streamlit,KubernetesAchievements:* Build highly scalable NLU inference platform to handle more than 100.000 messages per hour* Build next generation NLU platform supporting intent classification and named entity recognition* Introduced IaaC and CI/CD to the AI processes and the whole MLOps cycle* Build quick data prototypes for Customer Success & Sales Team * Delivered highly successful innovative ecommerce chatbot use cases (https://certainly.io/customers/tiger-of-sweden/)* Built a successful data analytics tool to prevent customer churning, saving the company half million on revenue.* Build tools to provide access to AI in a NoCode environment* Migration of an old NLU engine to faster, better and feature-rich NLU engine that allows to support 150.000 models in parallel

* AI solutions provider for medical sector (San Francisco, 100% Remote from Germany)* Responsible to put research into highly scalable and faults tolerant production environment* Advising company on technical decision and helping research scientists to automate their work* Tech: Python, FastAPI, Keras, TensorFlow, PyTorch, Cython, Qt, Docker, Kaldi, AWS, SQS, EC2, EKS, Terraform, ECS, Fargate, AWS Labmda* Migration of a monolithic C++/Qt backend to a scalable microservice based backend on AWS (SQS, S3, ECS, Fargate, EC2, Route 53, RDS, AWS Lambda, AWS Step Functions, Sentry, Docker, Terraform).* Building Speech Recognition Engine on top of kaldi with cloud native technologies. * Decoding more 10.000 audio hours a day. Still running in UK.

* Conception and implementation of data-driven projects for customers (50+ projects)* Presales activities for machine learning and NLP customer use cases* Implementation of cloud native projects on AWS and GCloud* Co-establishment of the data team: included the creation of offerings formachine learning and smart data, strategic alignment, direct customer contact and presales activities through brown paper bag lunches* Conducting workshops on ML and MLOps at various conference and at customers' sites* Implementation of remote workshops* Organization and development of the internal data science community

I wrote my bachelor thesis about text classification. Therefore, I use natural language processing and deep learning to accomplish this task. Incoming requests from the customers are pre classified for the customer service in order to solve them quicker.

I was part of a data mining project to retrieve insights about potential reinvestors in order to make customers specialised offers. The project was implemented with SPSS Modeler and python. Several models were developed.

Data analysis in order to get the preferred medium to contact the customer. I used scala, apache spark and apache zeppelin in an hadoop environment for the data analysis.

Together with another intern, I developed a fullstack application to visualize network components for IT/software architects. The server was developed with and the spring framework. The frontend was developed with web technologies. Since it was an agile project, we used methods like scrum and test driven development.

Developed a lot in Java and learnt about multi-threading. As my final project I developed a server-client application for drawing images together with colleagues across the network.

Early stage startup which scaled its app from 0 to 100.000 users in one month. I was responsible for backend development and data analytics.Tech: JavaScript, Node.js, Docker, Selenium, Ionic, Feathers.js, Azure