-
Fully declarative, feature-rich, deployable, modular and extensible.
/ ML Engineer // Rust & Python /
-
Joined on
2026-01-27
Hello, world! [status: building]
About Me
My main languages are Python and Rust. I build microservices, event-driven and ML pipelines. Fan of self-hosted services, experimenting with LLM agents (Harness Systems), and reproducible NixOS infrastructure. Getting into Machine Learning.
Stack
Languages: Python, Rust, Nix, SQL, Bash
ML & AI: LangGraph, Whisper, MediaPipe, spaCy, scikit-learn, NumPy, HuggingFace Transformers
Backend & Infra: FastAPI, Celery, NATS, PostgreSQL, Redis, SeaweedFS, Aiogram, Django
DevOps & Tools: Docker, Kubernetes (Kustomize), GitHub Actions, Prometheus, Grafana, Linux (Ubuntu, NixOS)
Pinned Projects
charisma-master - AI speech analysis platform: transcription and evaluation of presentations. 6 microservices on K8s. FastAPI, Rust, K8s, NATS, Whisper, LangGraph
magic-market - Production-grade marketplace with hardened K8s configuration. FastAPI, Rust, K8s, Docker, NATS
homelab-bot - Matrix bot for infrastructure management: Forgejo webhook integration, activity digests, RSS → LLM TL;DR. Rust, Matrix SDK, PostgreSQL, Docker
nix - Declarative NixOS config for 2 hosts (homelab + PC) with Flakes, age-encryption, and full self-hosted stack (Nextcloud, Forgejo, Matrix, Prometheus). NixOS, Flakes, home-manager, sops-nix
Currently
Diving into ML fundamentals - NLP, model math and implementations from scratch, model training competitions. Experimenting with LLM agents: tool-calling, multi-agent orchestration, structured outputs.
Collaboration
My Forgejo instance is private (homelab infrastructure), but I'm open to PRs, issues, and collaborating on open-source - especially Rust, Python infrastructure, or ML tooling. Best way to reach me is email (other contacts listed below).
Connect
