Leonard
Morgenschweis

I'm a Master's computer science student at TUM in Munich. I like building software that runs locally on the device, mostly small tools and experiments around AI and audio. A few of my public projects are below.

Projects

Aitvaras

Local-first AI assistant for macOS

I built Aitvaras after finding OpenClaw too operationally heavy for the desktop assistant I wanted: a focused native app that makes everyday computer work smarter. It combines fully local inference, bilingual voice interaction, inspectable long-term memory and typed macOS integrations with a playful 3D companion overlay.

Swift 6 · SwiftUI · SceneKit · MLX · On-device LLMs · GRDB / SQLite · EventKit

System overview

Interfaces

Core

Connectors

Inference and state stay on-device. Click a module to see what it does.

View on GitHub ↗

Hearing Aid

Software hearing aid in Rust

An experimental software hearing aid written in Rust. It runs a few common hearing-aid DSP steps (multiband compression, frequency shifting and noise suppression) through a low-latency audio pipeline, with optional live transcription via Whisper. Mostly a project to learn real-time audio processing.

Rust · DSP · cpal · RustFFT · whisper-rs · egui

View on GitHub ↗

Background

since 2026 M.Sc. Computer Science at the Technical University of Munich.
2022–2026 B.Sc. Informatik at the Technical University of Munich, minor in mechanical engineering. Thesis on a simulation environment for autonomous construction robots.
2025 Exchange semester at IIT Bombay: machine learning, blockchains, embedded systems.
since 2025 Working student (automation engineer) at Noyes Technologies, Munich.
2023–24 Working student at Infineon Technologies: graphics & embedded software for automotive human-interface devices, on evaluation boards and production hardware.