arxiv:2610.34594
🤝 Open to Collab
Mohsen Seyedkazemi Ardebili
mskazemi
AI & ML interests
AI Platform & Agentic Systems Engineer | Open-Source Developer & Researcher. Ph.D., University of Bologna. I build reliable, privacy-conscious AI systems, from agentic infrastructure and HPC automation to offline speech recognition and accessible human-computer interaction. Projects: KubeIntellect, a human-governed AI SRE for Kubernetes, peer-reviewed in the Journal of Grid Computing; NovaFabric, tamper-evident, replayable evidence for AI agent runs; YazSes, open-source offline voice typing for Linux, macOS and Windows. Interests: agentic AI, MLOps, reproducible evaluation, AI safety, speech recognition and human-centered computing. Education: Ph.D. in Electronics, Telecommunications, and Information Technologies Engineering, University of Bologna. Website: https://mskazemi.com/ | GitHub: https://github.com/MSKazemi
Recent Activity
posted an update about 12 hours ago
YazSes: offline voice typing built on faster-whisper (CTranslate2).
Hold a key, speak, release, and the text is typed into the focused app on Linux, macOS and Windows. Recognition runs on your own CPU, so by default no audio leaves the machine. It also transcribes recordings with optional speaker labels (`yazses transcribe interview.m4a`).
Try the transcription path in the browser (the hosted demo runs on a Space for convenience; the app itself runs locally): https://huggingface.co/spaces/mskazemi/yazses-offline-transcription-demo
Code and install options: https://github.com/MSKazemi/yazses (Apache-2.0) updated a collection 9 days ago
Agentic AI, HPC & Speech — Mohsen Seyedkazemi Ardebili updated a Space 9 days ago
mskazemi/kubeintellectOrganizations
