Skip to content
View richardharnisch's full-sized avatar

Highlights

  • Pro

Block or report richardharnisch

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.

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
richardharnisch/README.md

Hi, I'm Richard 👋

I'm an AI builder based in the Netherlands. In September 2026, I'll start the MSc Artificial Intelligence and a research assistant position at the University of Amsterdam, where I'll work with Dr. Shashikant Ilager on energy-efficient AI data-center design. I recently completed my BSc Artificial Intelligence at the University of Groningen cum laude (8.4/10).

I'm especially interested in agent harnesses, personal and business assistants, and reliable LLM systems. I also have industry ML experience from projects with Gasunie and descript.

Selected work

🥇 Maverx Style Buddy: agentic slide generation

First place at the Digilize AI Hackathon 2026. The system learns a visual style from existing presentations, then generates new, fully editable PowerPoint decks and supporting materials. My work covered the prompts, agent architecture, product framing, demo, and pitch.

My bachelor's thesis tested for double descent in DQN and TRPO and, unfortunately, reported a negative result. I am currently extending this work as I haven't given up on DD existing in RL yet.

Worked in a five-person team on forecasting Dutch electricity prices for Gasunie, with a focus on evaluation, retraining, and distribution shift rather than only in-sample model performance.

More things I've built
  • Clark Bott: my personal Hermes agent for managing tasks, coordinating my schedule, and running recurring briefing and research jobs.
  • Neural network from scratch: a CNN implemented with NumPy and tested on GTSRB.
  • PDF Prompt Injector: a small demonstration of invisible-text prompt injection in PDF documents.
  • Picmash: a Django-based Elo rating system for ranking photo-contest entries.
  • CoverRoomBets: a portfolio rebalancing utility that won Cover's trading competition.
  • Numista API interface: a tool I built for my dad to estimate the value of a numismatic collection.
  • TeX workshop: materials for a workshop I gave to chemistry students preparing to write their bachelor's theses.
  • Programming puzzles: Advent of Code and Crash & Compile, where I placed third overall and second in coding at Cover's event.

Contact

LinkedIn · CV · richitdd at Gmail

I'm always excited to hear from ambitious people and about hard technical problems and opportunities to do good.

Pinned Loading

  1. pytorch/vision pytorch/vision Public

    Datasets, Transforms and Models specific to Computer Vision

    Python 17.9k 7.2k

  2. rl_double_descent rl_double_descent Public

    Python 1

  3. maverx-style-buddy maverx-style-buddy Public

    Decker: Agentic Slide Deck Generator. Automatically ingests corporate style from pptx files and generates editable pptx.

    Python 2 2

  4. kodeInInk/PriceForecasting-GasuniexRUG kodeInInk/PriceForecasting-GasuniexRUG Public

    Jupyter Notebook 1

  5. coverroombets coverroombets Public

    Utility script for creating and rebalancing a market-cap weighted index portfolio for Cover's live simulated trading competition. Won first place!

    Python

  6. TeunB6/single-tree-detection TeunB6/single-tree-detection Public

    Python 1