Ram Maheshwari Logo Image
Afra Azad

Hello, My name is Afra

I am a Data Science student passionate about machine learning, LLMs, and NLP, with experience applying AI in healthcare and finance.

About Me I am a part-time final-year student in Data Science at the University of Toronto, St. George, and I plan to finish my studies at the end of 2026.

Get to know me!

Hey! I'm Afra Azad, a Data Science student at the University of Toronto with experience in machine learning, NLP, and data engineering across healthcare and industry.

I currently work as a Machine Learning Developer at UHN, where I build and deploy supervised learning pipelines, including LSTMs, ensemble models, and production SQL-backed prediction systems. Previously, I worked at the U of T Temerty Faculty of Medicine, where I applied transformer models (BERT, LLaMA) for NLP tasks, and at Culligan Quench engineering data pipelines and Power BI dashboards. Feel free view my resume here.

Resume

My Skills

Python
Machine Learning
LLMs & NLP
BERT / LLaMA
TensorFlow
Scikit-learn
HuggingFace
SQL / PostgreSQL
Power BI
Tableau
Google Cloud
MS Azure
Git
SAS

Projects A selection of my data science and machine learning work spanning healthcare, NLP, and data engineering.

NVDA Automatic Image Caption

This project focused on developing an AI-powered image captioning system for NVDA (NonVisual Desktop Access) to improve accessibility for individuals with visual impairments. The work involved researching existing machine learning approaches, processing and training on large-scale image–caption datasets, and building deep learning models using CNNs to generate accurate image captions. The project's impact lies in making digital content more accessible by helping users better understand and interact with images through automatically generated descriptions.

InboxIntel

This project, InboxIntel, focused on developing a machine learning system to automatically detect spam emails and identify the keywords most indicative of malicious or fraudulent content. Using a large-scale dataset of over 82,000 emails collected from multiple public corpora, the project involved cleaning and analyzing email text data, applying natural language processing techniques, and training predictive models including Naive Bayes, Logistic Regression, and Neural Networks. The project's impact lies in strengthening cybersecurity efforts by improving automated spam detection and helping protect users from phishing attempts, fraud, and other malicious online threats.

Summify

This project, Summify, focused on developing an AI-powered platform that automatically extracts and synthesizes information from audio and video content to save users time and improve accessibility through API calls. The system was designed to generate concise summaries, identify key points, extract major topics, and perform sentiment analysis across various audio and video formats. The project's impact extends beyond productivity by improving accessibility for individuals with hearing impairments, making information from audio and visual media easier to consume and understand.

Credentials & Languages Certifications I've earned and languages I speak.

Certificates

  • Learn to Program: The Fundamentals University of Toronto
    2022

Languages

  • English Native
  • Bangla Professional
  • French Classroom
  • Arabic Beginner

Contact Feel free to reach out! I am open to dicussing new opportunities, projects, or collaborations. You can also find me on LinkedIn.