Thanks to visit codestin.com
Credit goes to github.com

Skip to content
View ialwayslikedgrime's full-sized avatar

Block or report ialwayslikedgrime

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 supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
ialwayslikedgrime/README.md

👋 Hi, I'm Stella

Welcome to my space on GitHub.

I'm a passionate AI student from Northern Italy, currently enrolled at the Mathematik Fakultät at Humboldt Universität zu Berlin for the full 2025/2026 academic year (fully taught in German). I am an erasmus student from the Artificial Intelligence course at University of Milan.

Current Projects

  • Multi-Input Multi-Output Neural Network – Designed and implemented a multi-modal sentiment analysis system for hotel reviews.

    • Combined NLP (LSTM branch) with structured metadata (categorical branch) using the Keras Functional API.
    • Tackled multi-task learning: simultaneous classification (positive/negative) and regression (review score prediction).
    • Built a full preprocessing pipeline (text cleaning, tokenization, categorical encoding, temporal/geographic features).
    • Explored hyperparameter optimization (random search over learning rates, dropout, architecture size) and achieved robust performance across tasks.
    • Repo includes clear documentation of design choices, architecture visualization, and notebook implementation.
  • Airbnb Price Prediction – End-to-end machine learning pipeline using real-world Airbnb data (23K+ listings across 88 neighborhoods) with comprehensive data cleaning, feature engineering on classic tabular features and geospatial engineering with GeoPandas (neighborhood encoding, spatial joins), EDA, and model selection, optimization and validation. Interpretability through SHAP to surface business drivers (location, property type, amenities). Test performance XGBoost: R² = 0.587. Used software engineering principles applied to data science through modular code and clear folder organization (inspired by cookiecutter data science). Libraries: pandas • NumPy • scikit-learn • XGBoost • GeoPandas • Shapely • matplotlib • SHAP

  • Particle Swarm Optimization Particle Swarm Optimization (PSO) – From-scratch implementation of a swarm intelligence algorithm, part of the broader family of evolutionary and genetic-inspired optimization techniques. Vectorized with NumPy and engineered with OOP principles.

  • Whisper Transcription App – A Next.js web application that integrates Hugging Face Transformers with OpenAI Whisper for local AI transcription.

    • Built to streamline my study workflow by wrapping Python AI scripts in a user-friendly web interface.
    • Handles YouTube URLs, audio file uploads, and live microphone recording with real-time progress tracking.
    • Technical stack: Next.js, TypeScript, React, Tailwind CSS, Node.js, Python, Hugging Face Transformers, PyTorch for local Whisper inference, yt-dlp for YouTube processing, FFmpeg for audio handling.
    • Currently learning full-stack deployment: scaling from local execution to cloud architecture with user authentication (Clerk), database integration (PostgreSQL), and payment processing through Stripe API.

online ialwayslikedgrime , my real name is Stella Andorno

Pinned Loading

  1. ialwayslikedgrime ialwayslikedgrime Public

  2. multimodal-sentiment-analysis multimodal-sentiment-analysis Public

    Multi-input, multi-output neural network for hotel review sentiment analysis designed from scratch. Combines LSTM text processing with categorical feature engineering to predict both review sentime…

    Jupyter Notebook 1

  3. airbnb-milan-price-prediction airbnb-milan-price-prediction Public

    End-to-end machine learning pipeline for predicting Airbnb listing prices in Milan using geospatial analysis, nested cross-validation, and feature engineering. Integrates multiple data sources with…

    Jupyter Notebook 1 1

  4. Particle_Swarm_Optimization Particle_Swarm_Optimization Public

    Particle Swarm Optimization implementation demonstrating proper software engineering practices. Built as a consistent extension of the Softpy library using inheritance from the library's abstract b…

    Python

  5. transcription-app-prototype transcription-app-prototype Public

    Transcription app prototype built with OpenAI Whisper

    TypeScript