Abdulmunim Jundurahman

Developed an end-to-end Agentic HR System featuring an LLM-powered interviewer with a 3D avatar and voice interface. Integrated multi-modal sentiment analysis and a multi-agent AI screener, enabling HR to efficiently process over 10,000 candidates monthly.

Achieved a 15% improvement in dialect detection accuracy by fine-tuning speech recognition models on a custom dataset. Optimized the learn-quiz-reward model, driving a 125% increase in average session duration among teenage users.

Designed and implemented a Speech Recognition Model optimized for edge devices, capable of recognizing commands in multiple languages. Developed a custom speech-text alignment technique achieving 95% accuracy.

Explored NeuroSymbolic AI algorithms, contributing to research aimed at advancing the development of a general artificial intelligence system.

Developed emotion quantification models achieving 92% accuracy for various settings. Deployed models on edge devices to ensure data privacy and reduce latency.

Developed Low-Resource Language Translation Models for Amharic and Afan Oromo, achieving significant improvements in translation accuracy given the limited resources.

Grade: CGPA 3.9/4.0

BSc. in Software Engineering with a specialization in AI. Achieved a CGPA of 3.9/4.0, ranking among the top 10 universities in Africa.