Skip to content
HN

Hussain
Nasir

Software Engineer · AI / ML

I build intelligent systems that work in the real world, from healthcare workflows to autonomous drones.

▌

Hello, I'm Hussain.

I'm a Computer Science and Mathematics student at the University of Waterloo who likes turning messy, real-world problems into reliable software.

I've shipped LLM pipelines into production underwriting workflows, co-founded a healthcare workflow platform running in pilot clinics, and written computer-vision code that helps a UAV land on its own.

Education

University of Waterloo

Bachelor's (Honours), Computer Science and Mathematics (Co-op)

Graduation May 2028

Coursework

Data Structures & Algorithms · Linear Algebra · Statistics · Object-Oriented Programming

CS & Math (Co-op) @ University of Waterloo, graduating May 2029

97%+

recall on risk classification model

2

pilot clinics running MedFlow AI

30+ FPS

live UAV camera processing

(02) Experience

Where I've worked

Jan 2026 – Aug 2026 · Toronto, ON

Northbridge Financial Corporation

AI Engineer Intern

  • Built a risk classification model with 97%+ recall at 65%+ precision to minimize missed high-risk cases, evaluated across 5K+ underwriting submissions, with borderline cases routed to human underwriter review
  • Engineered production LLM APIs with structured prompts, JSON validation and fallback logic, achieving 99%+ valid structured outputs
  • Batched LLM prompt calls to cut per-submission latency and API costs
  • Deployed the risk classification pipeline as a FastAPI service for production underwriting workflows

(03) Products

Products I'm building

Real-world platforms, deployed and in active development.

01

Active development

AI Healthcare Workflow Platform

Co-founder / CTO · Mar 2026 – Present

A full-stack platform that automates clinic administrative workflows, deployed to pilot clinics and cutting administrative workload by roughly 50%.

  • LLM document-processing pipeline using structured extraction with 90%+ field-level accuracy
  • React + FastAPI apps with role-based dashboards, approval workflows and automated task orchestration
  • EMR integrations via APIs to auto-create patient records and sync downstream workflow data
ReactFastAPIPythonLLMsEMR APIs
Links coming soon

02

Active development

Cogd.ai

AI/ML-powered digital cognitive health dashboard · Founder

Uses keystroke dynamics to assess Parkinson's disease risk through a real-time ML inference pipeline.

  • Integrates 5+ ML models behind a real-time inference pipeline
  • Behavioral data pipeline capturing 7+ keystroke metrics and extracting 30+ statistical and temporal features
  • Trained and evaluated Random Forest, XGBoost, LightGBM, Logistic Regression and SVM on the Tappy dataset (Accuracy, Precision, Recall, F1, ROC-AUC)
  • Deployed React/FastAPI app with Supabase auth; users take typing assessments and receive generated reports
PythonFastAPIReactSupabaseVercelscikit-learn
Links coming soon

(04) Projects

Other Projects I've built

01

Autonomous UAV Landing

University of Waterloo Aerial Robotics Group · Sep – Dec 2025

Computer-vision landing-zone prediction integrated with the flight stack for end-to-end autonomous landing.

  • Processes live camera feeds at 30+ FPS
  • Landing-zone prediction at 92% accuracy across 5,000+ real-time frames
  • Integrated with the existing flight stack for end-to-end autonomous landing
PythonPyTorchNumPyOpenCV
GitHub
02

Chronic Kidney Disease Predictor

Machine learning · Healthcare

Uses a Random Forest Classifier optimized with GridSearchCV to predict CKD based on six clinical features: Age, BUN, Diabetes, Creatinine Level, Hypertension, and GFR.

scikit-learnRandom ForestGridSearchCV
GitHub
03

Sign Language Detector

Computer vision · Real-time detection

A real-time sign language detection system with 100% accuracy across all 26 different gestures, using YOLOv8 object detection to identify hand gestures.

PythonYOLOv8
GitHub
04

Semantic Image Search Engine

Multimodal AI · Retrieval

Multimodal AI image search engine using OpenCLIP and FAISS for natural-language retrieval across image datasets.

OpenCLIPFAISS
GitHub

(05) Skills

Toolbox

Languages

Python / C++ / C / JavaScript / SQL / Bash / HTML/CSS

AI / ML

PyTorch / scikit-learn / YOLOv8 / OpenCV / MediaPipe / LLM pipelines / Structured extraction

Data

pandas / NumPy / Matplotlib

Backend & Cloud

FastAPI / REST APIs / Microsoft Azure / Railway / Supabase

Tools

Git/GitHub / CI/CD / Linux

Microsoft Azure Fundamentals (AZ-900)