Skip to content
View rafallex's full-sized avatar
🎯
Focusing
🎯
Focusing
1 follower · 2 following

Highlights

  • Pro

Block or report rafallex

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
rafallex/README.md

Hi, I'm Rafael

M.Sc. student in Image Analysis and Machine Learning at Uppsala University (Sep 2025 – Jun 2027). Previously B.Sc. in Data Science at NOVA IMS (Lisbon) with an Erasmus semester at Lund University. Portuguese, based between Stockholm and Uppsala.

What I've been working on

  • LoRA fine-tuning of Qwen3-0.6B for a Greek mythology chatbot — SFT on a BeautifulSoup-scraped Theoi.com corpus with generic Q&A mixed in to prevent catastrophic forgetting; Gradio UI for side-by-side base vs fine-tuned comparison. → greek-mythology-chatbot-1RT730
  • Vision Transformers for emotion recognition — fine-tuned ViT-Base on RAF-DB (96% test accuracy on 7-class emotion) and served it via FastAPI behind a Furhat social robot tutor. → intelligent-interactive-systems-1MD032
  • Changepoint / HMM methods for microfluidic bacterial growth — benchmarked five time-to-detection methods on real microfluidic data. → research-methodology-1MD048
  • Deep learning from scratch — NumPy-only fully-connected network to residual CNNs on MNIST (98.0% → 99.29% test accuracy). → deep-learning-1RT720

Tools I reach for

Languages: Python (daily), SQL, R, MATLAB, Bash ML / DL: PyTorch, Hugging Face Transformers, PEFT / LoRA, TRL, scikit-learn, NumPy, Pandas, SciPy Stats & applied math: Gaussian Processes, Hidden Markov Models, Bayesian inference, changepoint detection, bootstrap, functional data analysis Systems: Git, Docker, FastAPI, SQLite, Streamlit, Gradio

Currently

Open to machine learning engineering, data science, applied statistics, and research / PhD positions in Sweden and the EU — full-time from June 2027, summer 2026 internships now.

Reach me

Pinned Loading

  1. intelligent-interactive-systems-1MD032 intelligent-interactive-systems-1MD032 Public

    Forked from Danielbb14/InteligentSystemsProject

    Emotion-aware multimodal tutor pairing a Vision Transformer (fine-tuned on RAF-DB) with a Google Gemini dialogue layer, served via FastAPI to drive a Furhat social robot. Uppsala University, Intell…

    Jupyter Notebook

  2. Research-Methodology-for-Image-Analysis-and-Machine-Learning-1MD048 Research-Methodology-for-Image-Analysis-and-Machine-Learning-1MD048 Public template

    Benchmarking Gaussian processes, PELT changepoint detection, bootstrap, HMMs, and functional data analysis for detecting microbial growth onset in microfluidic data.

    Jupyter Notebook

  3. introduction-to-image-analysis-1MD110 introduction-to-image-analysis-1MD110 Public

    Image analysis labs in MATLAB — thresholding, segmentation, SIFT feature-based registration, intensity-based registration, and motion tracking. Uppsala University course work.

    MATLAB

  4. modelling-football-1RT001 modelling-football-1RT001 Public

    Mathematical Modelling of Football (1RT001) — Uppsala coursework: xDA-P pass model, pitch exploration, statistical testing.

    Jupyter Notebook

  5. deep-learning-1RT720 deep-learning-1RT720 Public

    NumPy and PyTorch implementations from Uppsala's Deep Learning course (1RT720): MNIST classifier built from scratch, plus a progression from FC to residual CNN with BN and Dropout.

    Jupyter Notebook