Yusuf Solomon

I’m a Machine Learning Engineer focused on building and deploying AI that performs reliably in production. I work across deep learning, transformer-style modeling, and end-to-end ML pipelines with a practical, systems-first mindset.

I’ve trained and monitored ML models in production, built inference pipelines using PostgreSQL, and fine-tuned Computer Vision models using TensorFlow and Keras—improving accuracy from 25% to 70%. I also develop dashboards that help teams monitor model performance and catch issues early.

My applied AI interests go beyond training models: I design inference logic for real applications (including implementing YOLO inference in Java for Android) and build backend services for tasks like image auditing. I enjoy the challenge of turning research ideas into scalable, maintainable products.

In personal projects and blogs, I’ve implemented transformer training pipelines from scratch, explored LLM multi-agent reasoning with prompt orchestration and conversational memory, and fine-tuned large language models using QLoRA. I’m excited by measurable outcomes—clean data, robust evaluation, and production monitoring that makes intelligence dependable.