cv
Basics
Name | Abdulmunim Jundurahman Jemal |
Label | AI / ML Engineer and Researcher |
abdulmunimjemal@gmail.com | |
Phone | +251 900729000 |
Url | https://abdulmunim.me |
Summary | Machine Learning Engineer with expertise in NLP, backend development, and speech processing. Passionate about AI research and applying technology to solve real-world challenges, particularly in underrepresented languages and edge computing. |
Work
-
2024.10 - 2024.11 AI Engineer
SMARTi Co.
Designed and deployed advanced speech data processing pipelines, improving efficiency in large-scale AI applications of IoT user command recognition and fine-tuning models for real-time performance.
- Deployed efficient speech recognition models for IoT applications.
-
2024.05 - 2024.12 Software Engineer
Amazethu
Developed NLP models for a speech-based language learning app, optimizing speech recognition and translation accuracy for improved user interaction.
- Improved user experience through optimized NLP systems.
-
2024.02 - 2024.05 AI Research Intern
Icog Labs
Engaged in Neuro-Symbolic AI research, translating theoretical concepts into executable algorithms, and contributing to Explainable AI methodologies.
- Contributed to Neuro-Symbolic AI research.
-
2023.06 - 2024.01 NLP Intern
Lesan.AI
Built scalable data collection and preprocessing pipelines, automated preprocessing for multilingual low-resource machine translation, and contributed to end-to-end system improvements in the backend.
- Improved low-resource multilingual translation systems.
Education
-
2021.09 - 2026.06 Addis Ababa, Ethiopia
Bachelor
Addis Ababa University
Software Engineering (AI Specialization)
- Applied Mathematics I / II / II
- Statistics and Probability
- Mathematics for AI
- Social Network Analysis
- Machine Learning
- Web and Mobile App Development
- Data Analysis
- Data Structures and Algorithms
- Object-Oriented Programming
Certificates
Database Design Specialization | ||
University of Michigan (Coursera) |
Computer Vision Basics | ||
The State University of New York (Coursera) |
Natural Language Processing Specialization | ||
DeepLearning.ai (Coursera) |
Machine Learning Specialization | ||
Stanford University (Coursera) |
Projects
- 2024.01 - 2024.01
Afan Oromo News Classification System
Researched and implemented an entire custom data collection and preprocessing pipeline for “Afan Oromo” - a low resource local language, achieving 96.5% accuracy, 0.95 precision, and 0.97 recall in a news classification task. Additionally, I published the dataset on Kaggle for public usage.
- 96.5% accuracy
- dataset on Kaggle
- 2023.10 - 2024.02
Amharic Context-Aware Spell Checker
Developed a first-of-its-kind “Context-Aware” spell checker for Amharic - a low-resource local language. Improving the weakness of previous systems that were dependent only on dictionary lookup.
- Context-Aware
- weakness of previous systems
Skills
Programming | |
Python | |
JavaScript | |
TypeScript | |
C / C++ | |
SQL |
Technologies | |
Docker | |
PostgreSQL | |
Kubernetes | |
Git Actions | |
VectorDB | |
Distributed Systems | |
NoSQL and SQL Database |
Libraries | |
Hugging Face | |
PyTorch | |
TensorFlow | |
AutoML | |
Scikit-learn | |
Pandas | |
NumPy | |
Matplotlib | |
Seaborn | |
NLTK | |
SpaCy | |
Gensim | |
FastAPI |
Languages
Amharic | |
Native |
Afan Oromo | |
Native |
English | |
Fluent |
Arabic | |
Intermediate |
Japanese | |
Basic |
Interests
AI Research | |
NLP | |
Edge Computing | |
Low-Resource Languages | |
Model Quantization | |
Medical AI |
Books | |
The Beginning of Guidance (Al-Ghazali) | |
The Beginning and The End (Ibn Kathir) | |
Sapiens (Noah Harari) | |
Crime and Punishment (Fyodor Dostoyevsky) | |
From Zero to One (Peter Theil) | |
The Alchemist (Paulo Coelho) |
Animes | |
Attack on Titan | |
Vinland Saga | |
Demon Slayer | |
Naruto | |
One Piece |
Publications
-
2024.01.01 Afan Oromo News Classification Using Machine Learning Techniques
ResearchGate
This publication explores the application of machine learning techniques for classifying news articles in Afan Oromo.
-
2023.12.01 Developing a Context-Aware Amharic Spell Checker Using N-Gram Models
ResearchGate
This research focuses on creating a context-aware spell checker for Amharic using N-Gram models.