Muhammed Abdulrasheed

I build and deploy machine learning systems that actually run in production, not prototypes.

Currently an ML/AI Engineer at Filly Coder, where I design and ship FastAPI-based ML microservices powering a live EdTech marketplace (Filly Tutor). I’ve built and deployed three core systems end-to-end:

An NLP-powered tutor classification engine (TF-IDF + cosine similarity, sub-1s latency, 95% precision) A multi-factor recommendation system that replaced naive keyword matching with structured scoring A real-time DistilBERT sentiment analysis service (150–180ms latency, production SLA, 85% test coverage)

My work sits at the intersection of machine learning, backend engineering, and system design, covering model development, API architecture, and production deployment.

I hold a First Class BSc in Physics from the University of Ilorin, which shapes how I approach ML: quantitatively, rigorously, and beyond surface-level library use. Before Filly Coder, I worked as a Graduate Assistant at Fountain University, applying ML across healthcare, agriculture, and IoT research.

Outside my primary role, I build and deploy end-to-end ML products:

MATRA: an AI-powered academic writing platform used by students to generate structured final-year projects

ChaseBTC: an ensemble deep learning trading system (LSTM + GRU + Conv1D, Sharpe Ratio 1.24)

Little Botanist: a 97-class plant classification system using EfficientNetB0

Hey Rodea: a speech analysis system using Whisper ASR with multi-dimensional scoring

Technically, I work across the full ML stack: FastAPI, PyTorch, TensorFlow, scikit-learn, HuggingFace Transformers, Docker, and LLM systems (RAG, prompt engineering, LLaMA-based APIs).