list of papers, code, and other resources

What is the SOTA technique for forecasting day-ahead and intraday market prices for electricity in Germany?

A project focused on forecasting solar photovoltaic (PV) power generation using regional microclimate data. Implements machine learning models like CatBoost, LightGBM, and XGBoost for predictions, leveraging environmental features like temperature, humidity, wind speed, and solar radiation.

Interface enabling use of ANI-style, and other NN-IPs in the Amber molecular dynamics software suite. Works with both Amber engines, sander and pmemd.

Code for Kazeev, N., Al-Maeeni, A.R., Romanov, I. et al. Sparse representation for machine learning the properties of defects in 2D materials. npj Comput Mater 9, 113 (2023).

Paper in Science and Technology for the Built Environment about the GEPIII Competition

In this section, predicting the energy efficiency of buildings with machine learning algorithms.

My solution to solve the second IEEE-CIS technical challenge

ConfRank+: Extending Conformer Ranking to Charged Molecules

Prediction of turbine energy yield (TEY) using Neural Networks

⚡ AI-powered energy consumption prediction app. Flutter + FastAPI + ML. Reduce bills & carbon footprint . Smart energy optimization! 🌱

Predicting the Energy consumed by appliances using Machine Learning algorithms built from scratch

Full-stack machine learning project for predicting building energy consumption using FastAPI, Next.js, and reproducible Jupyter notebooks.

Machine Learning Project on Electricity Consumption For Household Appliances. Random Forest gave us the best results. Model achieved 97% accuracy to optimize appliance‑level energy usage and reduce costs.

Wind Power Generation Forecasting using Machine Learning A data-driven forecasting project that uses machine learning models to predict wind power generation based on historical energy and weather data. Includes preprocessing, EDA, model training, and performance evaluation—all inside a well-documented Jupyter Notebook.

A Flask-based web application to forecast wind turbine renewable energy generation using time-series feature engineering and a pre-trained XGBoost model. Users can input custom date ranges and visualize future energy predictions through dynamic Matplotlib plots.

Deep learning regression project that predicts power plant energy output using an Artificial Neural Network (PyTorch) with live deployment via Streamlit.

SynapticGrid is an AI-driven system designed to make cities more efficient, sustainable, and livable by optimizing smart energy grids, waste management, and traffic flow through IoT sensors, real-time data processing, and reinforcement learning algorithms. The modular platform continuously learns and improves, helping urban environments

Pytorch implementation of Alchemical Kernels from Phys. Chem. Chem. Phys., 2018,20, 29661-29668

Time Series Forcasting and Clustering for Energy Management - Machine Learning & Imputation

Experimental data used to create regression models of appliances energy use in a low energy building.

Pipeline complet de prédiction de consommation électrique. Multiples modèles (ML, Deep Learning, séries temporelles) avec interface Streamlit et optimisation d'hyperparamètres.

Machine learning project predicting energy consumption in East Melbourne WWTP using ensemble and deep learning models, with SHAP-based feature importance analysis.

Predict household electricity consumption using a Flask API and Flutter app. Full-stack ML project for energy consumption forecasting based on voltage and sub-meter data.

Predicting electricity demand using LSTM and Random Forest models. A Comparative study with load & weather data

This Python project demonstrates real-world AI solutions across multiple domains: motion detection using OpenCV, environmental monitoring with anomaly detection, energy consumption forecasting, and predictive maintenance for machinery. The system integrates Streamlit dashboards for interactive visualization and allows users to upload custom dataset

This project addresses the challenge of predicting baseline energy usage in buildings for performance-based financing. Participants estimate what energy a building *would have used* without retrofits, enabling fair billing and encouraging investment in energy efficiency.

Ce projet vise à prédire la consommation énergétique des appareils électroménagers en s'appuyant sur diverses données environnementales afin de réduire la consommation d'énergie et les émissions de carbone associées, contribuant ainsi à une gestion énergétique plus efficace.

This project is to develop a robust model capable of accurately predicting energy consumption in buildings. This endeavor involves harnessing historical energy usage data in conjunction with diverse weather and environmental variables to construct an effective predictive model.

This project explores two different investigation using methods of machine learning and hybrid approach to predict the peak energy consumption in Ireland and energy production in Portugal based on weather and time variables by evaluating performance for classification metrics and regression metrics.