Skip to content
View httpirsh's full-sized avatar
🎯
Focusing
🎯
Focusing

Highlights

  • Pro

Block or report httpirsh

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
httpirsh/README.md

Hi, I'm Íris 👋

Junior AI & Data Engineer · MSc student in AI & Data Science @ University of Coimbra

LinkedIn Email GitHub

I'm currently developing my thesis at INESC, mapping invasive alien plants using citizen science, satellite imagery, and earth observation foundation models — while also exploring job opportunities alongside it.

  • 🎓 MSc in AI & Data Science, University of Coimbra (2024 – Present)
  • 🌱 BSc in AI & Data Science, University of Coimbra (2021 – 2024)
  • 🛰️ Thesis @ INESC: invasive alien plant mapping with citizen science + earth observation foundation models
  • 📍 Coimbra, Portugal

🔧 Tech Stack

Programming & Tools

Python Java SQL Docker Kubernetes

AI & Machine Learning

PyTorch TensorFlow Scikit-learn Transformers LLMs VLMs

Data Engineering & Visualization

Kafka Spark Pandas NumPy PostgreSQL Matplotlib Superset

Geospatial & Earth Observation

QGIS GDAL Rasterio Remote Sensing

🚀 Projects

CO₂ Data Engineering & ML Pipeline Distributed pipeline (Kafka, Spark, PostgreSQL) with K-means clustering (Silhouette ≈0.84) to help stakeholders detect global CO₂ emission patterns and target mitigation efforts.

Image-to-Prompt Inversion Recovers high-quality prompts from images for controllable generation and dataset augmentation, combining VLM captioning, diffusion models, and CLIP/LPIPS-guided optimization (0.81–0.94 CLIP similarity).

ArtBench-10 Generative Modelling Compared VAE, DCGAN, and diffusion models on ArtBench-10 (FID/KID metrics) to guide model selection based on fidelity/diversity trade-offs.

🤝 Organisations

  • Núcleo de Estudantes de Informática, University of Coimbra — External Relations & Career Development Collaborator (2023–2024): managed company partnerships for the career fair and helped organize the first edition of UPtoData.
  • Portuguese Space Agency — "Terra em Foco" Event Staff (2026): supported event organization and logistics.

🌐 Languages

Portuguese (native) · English (fluent)

📊 GitHub Stats

Pinned Loading

  1. co2-emissions-kafka-spark-pipeline co2-emissions-kafka-spark-pipeline Public

    CO2 Data Engineering Pipeline with Kafka, Spark, PostgreSQL, and Superset on Kubernetes.

    Jupyter Notebook

  2. miacd-iag-image-to-prompt miacd-iag-image-to-prompt Public

    VLM-initialized, metric-guided prompt inversion — recovering text prompts that reproduce target images through a fixed Latent Consistency Model, scored by CLIP/LPIPS/MSE and refined via LLM feedback.

    Python

  3. miacd-iag-generative-modelling miacd-iag-generative-modelling Public

    Comparing VAE, cVAE, DCGAN, WGAN-GP, and Diffusion models trained from scratch on ArtBench-10, evaluated with FID/KID.

    Jupyter Notebook

  4. lecd-pln-question-answering lecd-pln-question-answering Public

    Aplicação de dois sistema de Question Answering, baseados em regras e em métodos de Aprendizagem Computacional

    Python