Project-25012   Weather Prediction with AI

Project Title    

Weather Prediction with AI

Alex  Liu,  Director,   RMDS Lab and SPT Energy AI Lab
Email    alex@rmdslab.com

Project Description    

This project, “Weather Prediction with AI”, focuses on building high-quality predictive models for short-term and long-term weather forecasting using advanced machine learning and artificial intelligence techniques. Interns will work on the complete lifecycle of model development—from data preprocessing and feature engineering to model selection, evaluation, and deployment.

A major emphasis will be placed on improving the performance and robustness of existing forecasting models, as well as experimenting with AI agents to automate components of the workflow, including data ingestion, model tuning, and validation processes. Interns will gain hands-on experience with spatiotemporal climate data and work with large-scale, real-world datasets to extract actionable insights and enhance forecasting accuracy.

This project is supervised by Dr. Alex Liu, former IBM Chief Data Scientist, renowned for his award-winning predictive models in climate analytics and AI-driven environmental forecasting. Interns will benefit from direct mentorship and exposure to cutting-edge methods in AI-driven geospatial modeling.

This is an excellent opportunity for students interested in the intersection of climate science, AI, and scalable data systems to contribute to meaningful and impactful research."
Tasks and Responsibilities    "Interns will be expected to contribute across all stages of AI model development for weather prediction. Key responsibilities include:

Data Collection and Processing: Accessing, cleaning, and integrating spatiotemporal datasets from public climate data sources such as ERA5, NOAA, and NASA.

Exploratory Data Analysis (EDA): Performing in-depth analysis to understand patterns, correlations, and data quality issues.

Feature Engineering: Designing and testing domain-relevant features to improve model accuracy and robustness.

Model Development and Benchmarking: Building and comparing multiple machine learning models (e.g., regression, random forest, LSTM, transformer-based models) for weather forecasting.

Workflow Automation: Assisting in the design and implementation of AI agents to automate repetitive tasks such as model training, hyperparameter tuning, and result summarization.

Performance Evaluation: Using statistical and machine learning metrics (e.g., RMSE, MAE, R²) to assess and compare model predictions.

Documentation and Reporting: Maintaining thorough documentation of code, data sources, and findings; preparing periodic progress updates.

Collaboration: Participating in weekly check-ins with Dr. Alex Liu and other team members to present progress, receive feedback, and iterate on solutions."

Minimum Qualifications     

To be considered for this internship, applicants must meet the following criteria:

Educational Background: Currently enrolled in or recently completed an undergraduate or graduate program in Computer Science, Data Science, Atmospheric Science, Environmental Engineering, Geography, or a related field.

Programming Skills: Proficiency in Python, including experience with libraries such as NumPy, Pandas, scikit-learn, and Matplotlib.

Machine Learning Knowledge: Foundational understanding of machine learning concepts and techniques, including regression, classification, model evaluation, and overfitting.

Data Handling Experience: Comfort working with large datasets, including tasks like data cleaning, transformation, and visualization.

Strong Analytical Thinking: Ability to approach complex problems methodically, ask meaningful questions, and interpret quantitative results.

Communication Skills: Clear written and verbal communication skills, with the ability to document findings and present progress to a technical audience.

Self-Motivation: Capacity to work independently and manage time effectively in a remote or hybrid research environment.
 

Terms of the Project
3 months or 6 months or one year

Key Words    

Artificial Intelligence  Machine Learning  Weather Forecasting  Climate Data  Remote Sensing  Spatial Analysis  Data Science  Time Series  Python  Geospatial Data

Research Topics    

Weather Forecasting Models: Exploring short-term and long-term weather prediction using AI/ML approaches.  Model Selection and Evaluation: Comparing traditional statistical models with machine learning models (e.g., Random Forest, LSTM, Transformer-based models) for predictive accuracy.  Feature Engineering for Climate Data: Designing relevant spatiotemporal features from meteorological and remote sensing datasets (e.g., temperature, pressure, wind speed, humidity).  Spatial-Temporal Data Analysis: Investigating how spatial dependencies and temporal dynamics impact prediction performance.  Data Integration and Preprocessing: Handling heterogeneous datasets, missing data, and standardizing input formats for modeling.  Model Validation and Performance Metrics: Utilizing techniques such as cross-validation, RMSE, MAE, and accuracy comparisons.  Explainability in AI Models: Applying SHAP, LIME, or similar tools to interpret model decisions in a climate context.

Deliverables     

By the end of the internship, each intern is expected to produce the following:

Working Predictive Models: At least one fully developed and validated weather prediction model using real-world climate datasets.

Feature Engineering Pipeline: A reusable and documented pipeline for preprocessing and feature extraction from raw spatiotemporal data.

Automated Workflow Components: Scripts or modules that demonstrate AI-assisted automation in tasks such as model selection, tuning, or data handling.

Performance Benchmark Report: A comparative analysis of different models, including metrics, strengths, limitations, and recommendations.

Project Documentation: Clear and reproducible codebase, with README files, notebooks, and data usage notes.

Final Presentation: A concise presentation summarizing the methodology, results, and future directions, to be delivered to the Spatial Data Lab team and invited guests.

Optional but encouraged:

Blog Post or Medium Article: A public-facing summary of the project’s goals, approach, and key findings, co-authored with the team.