- Built high-throughput ML pipelines and tried automating things.
- Reduced inference latency of ML-based email phishing detectors by 90%+ (which really asked for it).
- "survived" an exchange at the University of Basel in Switzerland
- Got a Master's in Software & Data Engineering
| Category | My "I've seen things" list |
|---|---|
| Machine Learning | Transformers, LLMs, PyTorch, Hugging Face, Knowledge Distillation |
| Production ML | ONNX, TensorRT, CUDA, GPU Acceleration, High-Throughput Pipelines |
| Data & Cloud | AWS (Sagemaker, EC2), Docker, Apache Spark, Redshift, RavenDB |
| Languages | Python (Primary), C++, C#, Java, JavaScript |
- Phishing Detection – My MSc Thesis. Distilling LLMs because GPUs are expensive and LLMs are not fast enough without modifications for streaming data.
- Swiss Rail Visualizer – Tracking 2,500+ Swiss train connections in real-time.
- ToTheMoon – A cryptocurrency trading bot (with C# binding as well).
