Yunusa Jibrin

AI & Machine Learning Engineer | LLMs, Multi-Agent Systems, Production AI

πŸ“ Newcastle, UK | πŸ‡¬πŸ‡§ Global Talent Visa(No Sponsorship required)

About Me

Yunusa Jibrin

Hi, I’m Yunusa! I love transforming ideas into AI-powered solutions that people can actually use. From analyzing images with Computer Vision, to building smart NLP systems, to deploying models in production, I thrive at the intersection of research and impact.

Lately, I’ve been diving into LLMs, RAG systems, and GPU acceleration, exploring ways to make AI faster, smarter, and more reliable in real-world applications.

Note: All projects on this site are personal portfolio projects. Metrics are derived from controlled experiments, benchmarks, and simulated workflows, not from paid client or company deployments.

Featured Projects

πŸ’Š Healthcare AI & Compliance

Pharma AI Multi-Agent System

End-to-end AI system for pharmaceutical sales that integrates a **multi-agent compliance layer** for detecting potential off-label and regulatory risks with a **separate LLM-driven advisory agent ** that generates compliant, context-aware sales guidance. The system orchestrates agents in real time, applies rule-based and prompt-level constraints to reduce hallucinations, and is evaluated using synthetic and curated datasets to simulate decision-making in regulated sales environments.

OpenAI GPT-4 FastAPI Next.js Multi-Agent AI Vercel Render
πŸ€– LLM & Multi-Agent AI

Contract Intelligence System

Multi-agent AI platform analyzing legal contracts, identifying risks, and generating executive-level recommendations in under 60 seconds.

LLMs CrewAI GPT-4o FastAPI AWS EC2 System Design
🧠 NLP & RAG

QwenRAG PDF Search

Production-grade RAG system for PDFs with automatic citations and fast retrieval of answers.

LangChain ChromaDB Qwen2.5
πŸš— MLOps & CV

YOLO Pothole Detection API

Real-time pothole detection API using YOLOv8, deployed on AWS ECS for scalable infrastructure.

YOLOv8 FastAPI Docker AWS Streamlit
🎭 Computer Vision

Multi-Task Facial Attribute Prediction

Progressive CNN achieving 84.9% gender accuracy and high-performance facial attribute recognition.

TensorFlow Computer Vision Deep Learning Streamlit

Contact Me

Let’s connect β€” I’m always open to opportunities and collaborations.