SEEKING WORK | Oslo, Norway (CET) | Remote (EU/US time zones)
ML/LLM and systems engineer working in Python/PyTorch and Rust. 8+ years of professional experience with Python and machine learning; 5 years of production Rust. PhD in math, MSc financial economics, MSc neuroscience
Available for freelance work on:
- performance: profiling Python services, identifying bottlenecks and moving performance-critical code into Rust w/ PyO3
- LLM systems: agentic workflows with tools, planning and evaluation; inference serving w/ vLLM and TensorRT; edge inference
- Backend and data systems/data engineering: async Rust services, databases and high-throughput data pipelines.
- Quant and finance: backtesting engines, financial risk and fraud modeling
- Image and audio: applied ML, including generative models
Recent Rust projects:
- Leone: LLM inference engine written from scratch with my own CUDA/Metal kernels and a OpenAI-style server (https://github.com/t0rsion/leone)
- Holos: persistent homology/topological data analysis library with SOTA speed (https://github.com/t0rsion/holos)
B2B invoicing through my own company
Stack: Rust (Tokio, Axum/Actix, PyO3, Polars/Arrow), Python (PyTorch, NumPy/Pandas, Django/Flask), vLLM, TensorRT, Ray, Postgres, ClickHouse, Qdrant, Redis, Kafka, Docker, Kubernetes (k8s), gRPC, CUDA and Metal kernels.
Email: hn (the-at-symbol) vixe (dot) re
ML/LLM and systems engineer working in Python/PyTorch and Rust. 8+ years of professional experience with Python and machine learning; 5 years of production Rust. PhD in math, MSc financial economics, MSc neuroscience
Available for freelance work on:
- performance: profiling Python services, identifying bottlenecks and moving performance-critical code into Rust w/ PyO3
- LLM systems: agentic workflows with tools, planning and evaluation; inference serving w/ vLLM and TensorRT; edge inference
- Backend and data systems/data engineering: async Rust services, databases and high-throughput data pipelines.
- Quant and finance: backtesting engines, financial risk and fraud modeling
- Image and audio: applied ML, including generative models
Recent Rust projects:
- Leone: LLM inference engine written from scratch with my own CUDA/Metal kernels and a OpenAI-style server (https://github.com/t0rsion/leone)
- Holos: persistent homology/topological data analysis library with SOTA speed (https://github.com/t0rsion/holos)
B2B invoicing through my own company
Stack: Rust (Tokio, Axum/Actix, PyO3, Polars/Arrow), Python (PyTorch, NumPy/Pandas, Django/Flask), vLLM, TensorRT, Ray, Postgres, ClickHouse, Qdrant, Redis, Kafka, Docker, Kubernetes (k8s), gRPC, CUDA and Metal kernels.
Email: hn (the-at-symbol) vixe (dot) re