Skip to content
View NicolasLasch's full-sized avatar
:shipit:
:shipit:

Highlights

  • Pro

Block or report NicolasLasch

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
NicolasLasch/readme.md
Nicolas Lasch — Useful systems. Unexpected worlds. An original pixel explorer stands beside a portal in a quiet, dotted landscape.

Nicolas Lasch

Software engineer & AI researcher. Based in Belgium. Master's degree in Design and Artificial Intelligence, SUTD.

I build at the intersection of software engineering, AI, and interaction design: local AI tools, machine learning services, and multiplayer worlds. I like turning complicated systems into something people can actually use—or play.

Explore my repositories · Find me on X

Selected work

Tidy · Bergen · Hospital DTwin · Generative Audio

01 / Tidy

A pixel illustration of the explorer at a file workshop, matching the profile banner.

A local AI assistant for your Mac's files.

Ask in plain English, review the exact plan, and undo changes. A Rust safety engine handles execution; the optional local model helps interpret requests.

Rust Tauri TypeScript llama.cpp

Explore the project → · Download

Watch the real app demo Real Tidy app demonstration: file requests, reviewable plans, and undo.

02 / Bergen Bike Analysis

A pixel illustration of the explorer riding a bicycle beside weather and forecast symbols. This is decorative artwork, not a data chart.

Bike-demand forecasting, beyond the notebook.

Historical usage, time, and weather feed a prediction pipeline. A Flask API, Docker container, and AWS deployment workflow connect the model to a usable service.

Python Machine learning Flask AWS

Try the web app → · Source

Open the real dashboard preview Real screenshot of Bergen Bike Intelligence's prediction timeline, showing its summer station analysis.

03 / Hospital DTwin

A pixel illustration of the explorer visiting an eight-bed digital ward and telemetry display. This is project artwork, not an app screenshot.

A hospital ward you can explore in 3D.

An educational digital twin with bed inspection, telemetry streams, and patient dashboards. Three.js connects spatial interaction with demonstration patient data.

JavaScript Three.js Socket.IO

Explore the project →

Inspect the architecture Source-based Hospital DTwin architecture: an eight-bed 3D ward linked to patient dashboards.

04 / Generative Audio

A pixel illustration of the explorer wearing headphones, with video frames and a gently changing waveform. Matching artwork for the generative audio project.

From silent video to generated sound.

An experimental pipeline combines visual scene analysis, LLM audio planning, AudioLDM2 generation, and CLAP scoring to create sound effects for video.

Python Vision-language models AudioLDM2

Explore the project →

Follow the audio pipeline Generative Audio pipeline: perception, LLM planning, AudioLDM2 generation, and CLAP scoring.

Also in the workshop

Project What I built
Character Clips YouTube clipping around recognized faces, using YOLOv8 and face recognition.
Karmine SMP A Forge multiplayer world with custom dimensions, portals, guild cards, and rank-based access.
Farkle Dice A Fabric dice game with villager opponents, betting, and player challenges.
FoliaMarket / UserShops Java plugins for server markets and player trading.
NGNLUHC A No Game No Life UHC mode with custom multiplayer systems; still in development.
Odoo Hackathon 4.2 Collaborative development with the Macos-Linux team.

Field notes

Field notes: based in Belgium, Master's graduate in Design and Artificial Intelligence from SUTD, with Java, Python, Rust, TypeScript, and C# in the toolkit.

Languages · Java, Python, Rust, TypeScript, C#
Across projects · Local inference, computer vision, model serving, desktop apps, game systems.

Open the terminal
package com.github.nicolaslasch;

// A small profile, in Java.
public record Profile(String name, String[] interests) {
    public static void main(String[] args) {
        var nicolas = new Profile("Nicolas Lasch", new String[] {
            "Useful tools", "Learning systems", "Unexpected worlds"
        });
        System.out.println("Ready to build: " + nicolas.name());
    }
}
nicolas@workshop:~$ java Profile.java
Ready to build: Nicolas Lasch
nicolas@workshop:~$ _

Curious about how something works? Open a repository. The interesting part is inside.

Pinned Loading

  1. Tidy Tidy Public

    Tell your Mac what to do with your files in plain English. Review the exact plan, approve, and undo anything. Fast, private, and fully local.

    Rust 6

  2. NGNLUHC NGNLUHC Public

    No Game No Life UHC game mode, 90% done, still needs some fixes and testing. But main systems are done

    Java 2

  3. Bergen_Bike_Analysis Bergen_Bike_Analysis Public

    Production-ready machine learning project that forecasts bike demand at Bergen stations using historical usage data, time features, and weather information.

    Jupyter Notebook 7

  4. Odoo-Hackathons-Macos-Linux/hackathon-4.2 Odoo-Hackathons-Macos-Linux/hackathon-4.2 Public

    JavaScript 1