First ML hire for fast-growing Austrian AI upskilling startup (500K ARR in the first year). Led development for their AI course creation project and shipped first version for beta users in two months.

Early hire for VC funded audio ML startup, led by OSS Capital. Led prototype product development for ML-based audio tools. Spearheaded training and fine-tuning projects for music generation models. One of two main Python developers for nendo, the open-source AI audio tool suite. Developed scalable backend and infrastructure for serving music generation models on GKE on an 8xA100 cluster for >5000 beta users.

Conducted research project in cooperation between TU Vienna and sea.ai. Implemented a self-supervised learning approach for robust IR feature extraction. Developed PyTorch-based training, evaluation, and visualization pipelines for dataset of >2M LWIR images. Reduced company’s object detection system FP rate by 80%.

Led development of technologies to visualize and automatically process floor plans using ML methods. Designed and implemented processing pipeline supporting customer growth to 50K EUR MRR. Set up and maintained infrastructure for training segmentation models. Implemented graphics pipeline for automating visualizations using Blender, enabling processing of up to 1500 floor plans per month. Managed operations of a 6-person engineering team.

Developed backend JavaEE apps running on internal OpenShift cluster and handled REST-API governance. Led development of JMeter-driven test automation for REST-APIs, reducing test execution time by 150% and improving test coverage. Developed and integrated an OpenAPI generator into the company-wide internal Java framework, increasing development speed for new services by 30% and improving API documentation quality.

Junior Software Developer in the migration team, working on database migrations for large-scale customers.