Skip to content
View mustafameh's full-sized avatar
7 followers · 9 following

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Shark

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Shark

Block or report mustafameh

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mustafameh/README.md
Typing SVG

About Me

Applied Scientist at Thomson Reuters, building tool-augmented multi-agent systems for legal AI. I work at the intersection of LLMs, NLP, and real-world product deployment — from fine-tuning models for contract analysis to designing adversarial evaluation pipelines for agentic solutions.

  • MSc Artificial Intelligence (Distinction) & BSc Computer Science — University of Nottingham
  • Background in RLHF, computer vision, and ML for medical diagnostics
  • Published researcher in ML-driven medical diagnosis

Featured Projects

Sherlock-AI   Demo

A transferable character emulation pipeline with Sherlock Holmes as the proof of concept.

  • Fine-tuned a quantized LLaMA model with LoRA, automating dialogue extraction and synthetic data generation from public domain texts
  • ReAct-style reasoning for multi-step deductive analysis with dual inference (OpenRouter API + local GGUF)
  • Interactive story mode with branching mysteries, streaming structured output, and configurable writing styles
  • 72.4% user preference over baseline in A/B testing

Hugging Face LoRA Flask vLLM VITS

CourseCompanion

A RAG-powered educational platform for AI teaching assistants.

  • Professors create AI assistants that auto-generate knowledge bases from course materials
  • Chat interfaces, FAQ system, and management dashboards
  • RAG with LangChain and Flair for context-aware responses
  • Deployed on DigitalOcean with Google Drive integration

LangChain Flask PostgreSQL Flair OAuth

DeepGaze   Demo

Early autism detection from webcam-based eye-tracking data.

  • PCA + sklearn classifiers and a custom Keras CNN with transfer learning reaching 92% accuracy
  • Modular Flask web app for data collection, visualization, prediction, and model retraining
  • Published: Springer, 2024 — Pre-diagnosis for ASD Using Eye-Tracking and ML Techniques

Keras OpenCV WebGazer.js Flask sklearn

More Projects

Tech Stack

Languages

Python JavaScript TypeScript SQL

ML / AI

PyTorch Hugging Face TensorFlow scikit-learn LangChain

Infrastructure & Tools

AWS Docker PostgreSQL Flask Git

Pinned Loading

  1. Sherlock-LLM Sherlock-LLM Public

    Jupyter Notebook

  2. Automatic-Diagnosis-of-Autism-using-various-DeepLearning-models-on-eye-tacking-scan-paths Automatic-Diagnosis-of-Autism-using-various-DeepLearning-models-on-eye-tacking-scan-paths Public

    Jupyter Notebook

  3. Diagnosis-Of-Parkinson-s-and-Schzophernia-from-EEG-using-ML-and-DeepLearning- Diagnosis-Of-Parkinson-s-and-Schzophernia-from-EEG-using-ML-and-DeepLearning- Public

    Exploring and building different ways to help diagnose Mental Diseases using Machine Learning Techniques

    Jupyter Notebook 1

  4. Zeyad-Emara/Intelligent-Water-Quality-Prediction-System Zeyad-Emara/Intelligent-Water-Quality-Prediction-System Public

    An intelligent water quality prediction system developed as part of a University project. This system uses machine learning algorithms to analyze and predict water quality based on various water in…

    Jupyter Notebook 5

  5. BrickDestry_JavaProject BrickDestry_JavaProject Public

    HTML