Niels Verleysen

LinkedIn Medium GitHub HuggingFace

About Me and This Page

Welcome to my personal webpage! 🖐 I am excited to share my portfolio of side-projects and insights with you. Currently, I hold the position of AI & Data Chapter Lead at Verhaert, an innovation consultancy company trusted by industry leaders in their respective fields. 🚀 My role primarily revolves around applied data science research and AI engineering for product development, encompassing both physical and digital products. My expertise extends beyond ML algorithms as well, I thoroughly enjoy exploring new domains and creating innovative solutions.

Here's a fun fact: The digital painting you see was generated using a stable diffusion model that I fine-tuned on pictures of myself! It's an intriguing process that I highly recommend trying out if you have the opportunity. 😍

Profile Image

Projects

Transformer in PyTorch

Transformer

In this repository, I implemented the transformer architecture from scratch in PyTorch.

Implementing PPO for Atari Games

During my studies on Proximal Policy Optimization (PPO), I implemented the algorithm from scratch and trained agents to play Atari games using the Gym environment. PPO is a popular reinforcement learning algorithm known for its stability and sample efficiency. This project allowed me to gain hands-on experience with PPO and understand its inner workings.

Reddit TLDR with PyTorch and HuggingFace

When working with transformers, HuggingFace is an indispensable platform. It offers a wide range of resources including datasets, pretrained models, courses, and more, specifically tailored for transformers. In my pursuit to dive deeper into transformers, I embarked on various projects utilizing HuggingFace to gain hands-on experience. The Reddit TLDR project is one of several projects I created to explore the capabilities of transformers by summarizing Reddit posts.

View All Projects

Blogs & Insights

Why we should be worried about the whole “Sentient AI” story

Sentient AI

An engineer at Google makes a claim that an AI has become sentient. This should have us worried, but not for the reasons you might think…

The Future of Energy Efficient AI Systems

Join me and my colleague Bart as we discuss the current issues of AI, particularly deep learning technologies, due to their high power consumption and associated costs and carbon emissions. We explore new techniques like resource-aware AI to address these challenges and pave the way for energy-efficient AI systems.

Transformers Demystified

Picture taken during the talk

As part of my role at Verhaert, I delivered an internal talk on transformers. The session began with a brief overview of neural networks before delving into the theory and implementation details of transformers. I explained the inner workings of transformers and their applications in various domains, including natural language processing and computer vision. Additionally, I discussed how we can leverage pretrained models from HuggingFace to enhance our projects for clients. It was an engaging session that explored the power and potential of transformers in the field of AI.

All Insights