=========================
Certifications & Projects
=========================

Certifications
--------------

.. raw:: html

   <details class="collapsible-section">
      <summary>
         <h3>Huggingface - Agent Certification</h3>
         <p class="summary-line">Remote | Apr 2025</p>
      </summary>
      <div class="details-content">

.. figure:: /_static/assets/personal/huggingface/huggingface_agent_certification.webp
   :align: center
   :alt: Agent Certification
   :width: 70%
   :figclass: logo

*   **Summary:** I completed the Huggingface Agent certification, which involvs building your own `Custom Code Agent <https://huggingface.co/spaces/DenDeen/GAIA-Agent/tree/main/>`_. 
*   **Solutions:** The project focused on creating an agent capable of performing various tasks from the `GAIA <https://huggingface.co/spaces/gaia-benchmark/leaderboard/>`_ benchmark. The code agent is capable of a variaty of tasks required for the benchmark including text generation, and audio processing.
*   **Tech Stack:** Python, Huggingface Smolagents, Gradio, LangChain

.. raw:: html

      </div>
   </details>
   <hr/>

Projects
--------

.. raw:: html

   <details class="collapsible-section">
      <summary>
         <h3>Roundnet Noorderkempen - Website</h3>
         <p class="summary-line">Turnhout, Belgium | Jun 2022 – Now</p>
      </summary>
      <div class="details-content">

*   **Summary:** I created and maintain a website for a local roundnet club in Turnhout. 
*   **Website:** `Roundnet Noorderkempen <https://roundnet-noorderkempen.com/>`_

.. raw:: html

      </div>
   </details>
   <hr/>

   <details class="collapsible-section">
      <summary>
         <h3>LStat Datathon - AI Artwork Generation</h3>
         <p class="summary-line">KU Leuven, Belgium | Feb 2023</p>
      </summary>
      <div class="details-content">

.. figure:: /_static/assets/personal/datathon.jpg
   :align: center
   :alt: Datathon Picture
   :width: 70%
   :figclass: logo

*   **Challenge:** Create a PoC with a dataset containing images of artworks and descriptions.
*   **Solution:** Developed a custom diffusion `model <https://www.linkedin.com/feed/update/urn:li:activity:7039243586722222080/>`_ to generate artwork based on textual prompts. The model was trained on a dataset of 10,000 images and descriptions, allowing it to create unique pieces of art.
*   **Implementation:** Used PyTorch to train a custom diffusion model on Kaggle GPUs. Developed a Streamlit application for generating artwork based on textual prompts.
*   **Achievement:** Secured the "Best Pitch" award for the project presentation and demonstration.
*   **Tech Stack:** Python, PyTorch, Diffusion Models, Streamlit, Kaggle.

.. raw:: html

      </div>
   </details>
   <hr/>

   <details class="collapsible-section">
      <summary>
         <h3>Project Week - Audio book advertisement for a Danish brewery</h3>
         <p class="summary-line">Kolding, Denmark | Apr 2020</p>
      </summary>
      <div class="details-content">

*   **Challenge:** Create a fun advertisement for a Danish brewery.
*   **Solution:** Developed a website with me narrating funny Viking mythology. 
*   **Implementation:** Used HTML, CSS, and JavaScript to create an interactive website and audio player.

.. raw:: html

      </div>
   </details>
   <hr/>

   <details class="collapsible-section">
      <summary>
         <h3>Project Week - Emotion detection system</h3>
         <p class="summary-line">Chandigarh, India | Feb 2019</p>
      </summary>
      <div class="details-content">

*   **Challenge:** Collaboration with European and Indian students to create a project.
*   **Solution:** Developed a system to detect emotions in real-time using a webcam. The system used three different models, one for detecting faces in images, one for detecting predetermined facial features, and one to classify the facial features into classifications such as happy, angry, and others.
*   **Implementation:** Used Python and OpenCV to develop the system. The system was trained on small public datasets and was able to detect emotions in real-time.

.. raw:: html

      </div>
   </details>
   <hr/>