Open to opportunities

Hassan Raza

@hassanraza16

Data Scientist and AI/ML Engineer building predictive, NLP, and real-time systems that turn complex data into decision-ready insights.

Pakistan

Message

What I'm looking for

I’m looking for a role where I can build production-minded AI/ML—predictive modeling, survival analysis, and LLM/NLP systems—while collaborating cross-functionally to improve interpretability, reduce errors, and deliver real-time decision support.

I’m a Data Scientist and AI/ML Engineer focused on turning complex data into models and risk scores people can actually use. Across clinical risk stratification, real-time sensor analytics, and LLM-powered document workflows, I prioritize accuracy, interpretability, and deployment-ready pipelines.

Most recently, I analyzed a clinical dataset of about 140k patients to build and test risk stratification models for predicting sudden cardiac death post-myocardial infarction. I developed and validated survival models (including Cox, competing risks, and gradient boosting), achieving a C-index of 0.79, and improved model interpretability by 15% through conversion into easily interpretable risk scores.

In parallel, I’ve built practical AI/ML systems for messy, high-volume data. I developed custom web scraping pipelines (BeautifulSoup, Playwright), generated client-ready datasets using LLMs for automated CSV outputs, and built OCR workflows to extract information from pdfs and handwritten files.

I also delivered RAG and real-time analytics solutions, including a RAG system using FAISS and LLMs across 800+ PDFs, and a Flask-based system processing 100+ real-time sensor readings per minute. My approach blends statistical rigor, feature engineering, and cross-functional collaboration to improve prediction accuracy, reduce false alarms, and accelerate model performance.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Umbizo

Oct 2025 - Present (8 months)

Built and validated risk stratification models on a clinical dataset of ~140k patients to predict sudden cardiac death post-myocardial infarction. Improved model interpretability by 15% by converting complex outputs into clinically usable risk scores and leading cohort harmonisation, multiple imputation, and variable selection.

Survival Analysis Cox Proportional Hazards Competing Risks Gradient Boosting Model Validation Clinical Data Analysis Imputation Feature Engineering data preprocessing

AI/ML Engineer

EdgeC

Dec 2025 - May 2026 (5 months)

Developed web scraping pipelines to collect and analyze data from 100+ websites using BeautifulSoup and Playwright. Built client-ready datasets by integrating LLMs for automated CSV outputs and created an OCR pipeline to extract information from PDFs and handwritten files.

Web Scraping BeautifulSoup Playwright Data Engineering Python Text Analysis CSV

Machine Learning Engineer

Expert System Solution

Oct 2025 - Nov 2025 (1 month)

Built a Retrieval-Augmented Generation (RAG) system using FAISS and LLMs to extract information from 800+ PDF documents. Implemented drug-category clustering that improved clustering efficiency by 20% and deployed the solution on Azure Analytics.

RAG Faiss Large Language Models Clustering Predictive Analytics Python Data Pipelines

Data Scientist

AtkinRealis

Jan 2025 - Feb 2025 (1 month)

Developed a Flask-based AI system to process 100+ real-time sensor readings per minute for nuclear plant technicians, improving decision-making speed. Achieved 98% data accuracy by building Python/SQL cleansing and standardisation scripts, reducing false alarms.

Flask Real time Data Processing SQL Data Cleaning Data Standardization Python Machine Learning

Data Scientist

University of Essex

Aug 2023 - Jan 2024 (5 months)

Reduced machine learning training time by 20% using a Python-based data cleaning script with Pandas and NumPy and visualized results with Matplotlib. Improved model accuracy by 7% by applying A/B testing and experiment design across five feature extraction methods and addressing key prediction error causes with engineering and product teams.

Python Pandas NumPy Matplotlib A B Testing Statistical Modeling Supervised Learning regression Feature Extraction Data Visualization

Education

Degrees, certifications, and relevant coursework

University of Essex

Master of Science (MSc) in Artificial Intelligence, Artificial Intelligence

Grade: Distinction

MSc in Artificial Intelligence (Grade: Distinction) at the University of Essex, focusing on data mining, statistical analysis, and machine learning model optimisation. Awarded a $150 prize for achieving the highest project grade in the department.

BeaconHouse National University

Bachelor of Science (BSc) in Software Engineering, Software Engineering

Grade: CGPA: 3.73/4 (93.25%)

Bachelor of Science in Software Engineering at BeaconHouse National University, graduating with top academic performance. Achieved CGPA 3.73/4 (93.25%), was the top-performing student in the graduating class, and received a gold medal.