Issac Zerihun

Innovative professional with expertise in promoting academic, career and social development programs. Possesses excellent communication, problem-solving and organisational skills. Trusted advisor goes extra mile to help individuals succeed.

Full Stack Development Project: The Global Climate Archive

• Led the development of the Global Climate Archive, a full-stack web application aimed at providing comprehensive climate data analysis spanning 175 years (worldwide data).
• Fully designed application’s frontend and backend in 3 weeks using software development best practices.
• Ensured intuitive UI design adhering to the 10 Nielsen Design Heuristics, enhancing user experience and accessibility.
• Engineered an efficient database system using SQL and Java for seamless data management and retrieval.

Machine Learning Project: Predicting Energy Use

• Implemented and refined XGBoost and LSTM models for accurate energy use prediction with hourly data readings that had a range of 6 months.
• Conducted rigorous data preprocessing and hyper-parameter tuning to optimise model performance.
• Analysed model strengths, limitations, and proposed continuous improvement strategies based on peer-reviewed studies.
• Evaluated models using RMSE metric; XGBoost model achieved *(make it relevant or remove) an RMSE of 0.70 after refinement.

Data Analytics Project: Wine Analysis

• Led a data modelling project focused on predicting wine quality using advanced, supervised machine learning techniques.
• Analysed complex datasets to identify key factors influencing wine quality with over 10000 instances.
• Using model analysis and explanatory data, provided useful insights for optimisation that increased model accuracy by 5%.