Sosna Worku Achamyeleh

Junior Machine Learning Researcher

I collaborated on a predictive modeling project that focused on forecasting the risk of Coronary Artery Disease (CAD) and Heart Failure using data from the National Health and Nutrition Examination Survey (NHANES). Leading a team of 25, I managed data cleaning and preprocessing tasks, ensuring the datasets were prepared for analysis. I performed feature engineering and developed machine learning models using advanced techniques to predict health risks, working closely with a cross-functional team to ensure the accuracy and reliability of the models. This work contributed to potential advancements in healthcare outcomes by providing insights into health risk prediction.

Junior Machine Learning Research Team Co-lead

I led an 8-member team of junior machine learning researchers and engineers on the "Industrial CO2 Emissions Prediction with Machine Learning" project. My responsibilities included conducting comprehensive literature reviews to establish the theoretical foundation and collaborating with a global team to co-author a research paper, which is set to be published in the Omdena Journal of AI.

Backend Developer — Intern

I developed backend systems and RESTful APIs using ASP .NET Core for the HakimHub website, which helped reduce response times to under 90 milliseconds. Additionally, by applying "Clean Architecture" and following "Test-Driven Development" principles, I achieved 100% test coverage for the project.

Threat Detection in Spoken Amharic Language

I created a real-time threat detection system for spoken Amharic, using natural language processing (NLP) to classify audio data into threat and non-threat categories. To achieve this, I utilized various feature extraction techniques, including Mel-frequency cepstral coefficients (MFCCs), Chroma Features, and Spectral Features. I also incorporated data augmentation methods like time stretching, pitch shifting, and noise addition to make the model more robust and adaptable. Throughout the project, I explored different acoustic models, such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Long Short-Term Memory (LSTM) networks, to find the best architecture for accurately identifying threats in spoken language.

HakimHub

I developed a hospital and doctor recommendation system driven by user-provided symptoms. This project involved implementing Natural Language Processing techniques to extract disease information from user inputs effectively. Working within a group of 12, I contributed to the backend development by completing 18% of the endpoints using ASP.NET Core. One of the key features I implemented was the “Hospital Availability” functionality, which verifies the real-time availability of hospitals, enhancing user experience and ensuring timely access to medical care.

Tutorial Giving Mobile App

Developed a mobile application designed to facilitate self-learning for students while providing a platform to connect with tutors. The application features a robust RESTful API built with Spring Boot, following the Model-View-Controller (MVC) architecture for efficient data management. Additionally, I completed the user interface (UI) and implemented BLoC state management for 20% of the mobile app using Flutter.

Addis Ababa University

Year 5

CGPA: 3.62/4.0 (Great Distinction)

Coursework: Natural Language Processing, Artificial Intelligence, Machine Learning and Big Data, Data Structures and Algorithms, Object Oriented Programming, Operating Systems, Computer Architecture, Relational Database Management Systems

A2SV

Website

A2SV - a U.S.-based non-profit organization dedicated to upskilling high-potential African students has successfully placed over 60 students at leading technology companies such as Google, LinkedIn, Bloomberg, Databricks, Amazon, and Square over the past five years