IEEE Published97.1% accuracy
CNN Urban Sound Classifier
A convolutional network that classifies urban sounds, outperforming Random Forest (91.4%) and SVM (88.6%) baselines.
AI/ML Engineer → LLM & RAG Systems
NajmaAbdulkarimNour
AI/ML Engineer with 3+ years of experience turning models into real products — LLM applications, retrieval-augmented generation pipelines, and machine learning systems running in production. I have shipped software for a nationwide Electronic Health Records platform in Somalia and integrated LLMs into UN operational workflows at IOM.
I am an IEEE-published researcher with two peer-reviewed papers on urban sound classification and crash hotspot detection, and I am currently completing an M.Sc. in Artificial Intelligence at Canadian University Dubai.
Currently open to AI Engineer, ML Engineer and Data Scientist roles in the UAE. Get in touch.
Experience
Where I have worked
Research labs, UN operations and national health systems — building AI and software that people depend on.
2026 — Present
AI/ML Research Engineer
Unkad Labs — AI Research Lab · Mogadishu
- Built a complete LangChain RAG system: document ingestion, FAISS vector search, LLM integration and live deployment.
- Evaluated 500+ LLM outputs for safety, accuracy and performance with structured testing and reporting frameworks.
- Fine-tuned BERT for text classification and deployed it as a REST API with real-time inference monitoring.
Apr 2025 — Dec 2025
AI & Data Analyst
IOM Somalia — UN Agency · Somalia
- Integrated LLM APIs into UN operational workflows, cutting reporting time by 35% for UK and Denmark government donors.
- Built KPI dashboards and data pipelines used by senior international stakeholders across Somalia operations.
Oct 2024 — Mar 2025
Software Engineer
Raad System NGO — UNICEF-funded, Ministry of Health partner · Somalia
- Shipped production features for an Electronic Health Records system running across 80 healthcare facilities.
- Built data pipelines, REST APIs and Docker deployments in a live production environment.
Dec 2022 — Present
Co-Founder & Technical Lead
Uuntuse — Digital Non-Profit · Remote
- Built and maintain AI platforms serving organisations across 12+ African countries — Python, REST APIs and full-stack deployment.
Selected projects
Things I have built
LLMProduction
LangChain RAG Pipeline
End-to-end retrieval-augmented generation: document processing, FAISS vector search, LLM API integration and performance monitoring.
Healthcare ML+40% rare-class recall
ECG Anomaly Detector
Cardiac anomaly detection pipeline using SMOTE resampling, with a full Random Forest vs LSTM evaluation.
IEEE PublishedDBSCAN
Crash Hotspot Detection
Identifies high-risk traffic zones in Dubai using DBSCAN clustering and severity-weighted scoring.
Research
Peer-reviewed publications
IEEE · 2026
Urban Noise Monitoring and Classification Using ML and Deep Learning
Nour, N. et al. — IEEE International Conference on Smart Mobility, Egypt
DOI: 10.1109/SM69703.2026.11614142
IEEE · 2026
Crash Hotspot Detection in Dubai using DBSCAN and Severity Weighting
Alblooshi, M., Nour, N. et al. — IEEE International Conference on Smart Mobility, Egypt
DOI: 10.1109/SM69703.2026.11614141
Toolkit
Skills
- LLM & AI Engineering
- PythonLangChainRAGClaude APIGPT-4GeminiAgentic AIFAISS
- Machine Learning
- PyTorchTensorFlowBERTCNNLSTMHugging FaceScikit-learn
- Data Science
- Anomaly DetectionDBSCANSQLETLPandasNumPyMatplotlib
- Engineering
- DockerREST APIsFastAPICI/CDAzureAWSONNXC#JavaScript
Education
Study
Jan 2026 — Present
M.Sc. Artificial Intelligence
Canadian University Dubai, UAE
Machine Learning · Deep Learning · NLP · LLM Engineering · Computer Vision · Agentic AI
2020 — 2024
B.Sc. Computer Science
SIMAD University, Somalia
Elected Chairman of the Faculty of Computer Science, representing 900+ students.
Recent articles
Writing
Notes on LLM engineering, machine learning and building AI systems that hold up in production.