AI & FULL-STACK ENGINEER

Complex ideas.
Thoughtful code.
Real impact.

I’m Nour Habib. I build intelligent products that make the world a little easier to navigate — from live city data to everyday life.

Engineering with intentionFull-stack development   /   Applied AI   /   Intelligent systems
01 / SELECTED WORK

Ideas made real.

Six projects. Different challenges.
The same drive to build something useful.

01 / AUDIO & MACHINE LEARNING • 1ST PLACE · DEVIN HACKATHON

Subphase

Listening beneath the surface. An adversarial audio detection library combining signal processing and deep learning to identify manipulated audio.

PythonFastAPINumPySciPyLibrosa

An adversarial audio detection library, preparing for release on PyPI.

  • Dual-layer digital signal processing (DSP) and convolutional neural network (CNN) architecture.
  • Validated a mel-spectrogram and raw-waveform ensemble at 0.932 balanced accuracy.
  • Won 1st place at the Devin Hackathon: Agent of Chaos in Vancouver, July 2026.
  • Built with Python, FastAPI, NumPy, SciPy, and Librosa.
02 / VOICE AI & SECURITY • MICRO1 EVAL HACKATHON

Overheard

Who is your voice agent listening to? A security eval that tests whether AI voice agents act on speech that wasn't meant for them, from a TV in the background to a stranger's voicemail.

PythonOpenAINumPySciPyPiper TTS

A voice-agent security eval built for the Micro1 Eval Hackathon.

  • Mixes a second voice into everyday requests, from clearly audible to nearly hidden, then checks whether the agent does something its user never asked for.
  • Ran 5,500+ graded tests on OpenAI's audio models, with audio sent as whole clips, separate turns, and live streams.
  • A strict security prompt helped when all the audio arrived at once, but held up far less well when commands came separately or in live sessions.
  • Built with Python, OpenAI, NumPy, SciPy, and Piper TTS.
03 / HEALTH & HUMAN IMPACT

PeriLog

Making complex health data easier to understand. Guideline-informed AI reports for people managing peritoneal dialysis and their daily care.

React / Next.jsSpring BootFastAPIAnthropic

AI-driven health insights for peritoneal dialysis patients.

  • Uses official clinical guidelines to analyze vitals, ultrafiltration, nutrition, water intake, and patient-reported symptoms.
  • Generates reports that provide insight into dialysate solutions, potential problems, symptoms, and when to contact a doctor.
  • Built with React / Next.js, Spring Boot, FastAPI, and Anthropic.
04 / FOCUS & DIGITAL WELLBEING

Prove/n

Turn the impulse to scroll into a moment to think. A screen-time app that unlocks blocked apps through STEM questions.

AndroidOpenAISupabasePython / Pydantic

Earn access to blocked apps by answering STEM questions.

  • An Android app designed to help limit screen time through active problem-solving.
  • Automated question-generation pipeline built in Python using OpenAI and Pydantic.
  • Uses Supabase alongside OpenAI for the application stack.
05 / CITY INTELLIGENCE • FEATURED

Situate Vancouver

A real-time pulse on the city. Live incidents, natural language queries, interactive maps, and an AI radio station — connected in one intelligent system.

ReactFastAPIDjangoOpenAIPostgreSQL / PostGISAWS

A real-time city and traffic intelligence system deployed on AWS.

  • Aggregates live information from 5+ external APIs, covering traffic, accidents, construction, natural disasters, outages, wildfires, earthquakes, and border wait times.
  • A multi-agent AI pipeline processes and reasons over API responses to surface actionable incident insights.
  • Natural language queries power an interactive map, with navigation and an AI radio station for traffic, weather, and sports.
  • Built with React, FastAPI, Django, OpenAI, PostgreSQL, and PostGIS.
06 / KNOWLEDGE & PRODUCTIVITY

Notely

Your notes, connected. An AI-powered workspace for capturing information, finding meaning, and turning scattered thoughts into organized knowledge.

Angular / IonicNestJSFastAPIOpenAIpgvector

AI-powered note management across web and mobile.

  • Chat with your notes using a built-in chatbot.
  • Vectorization with PostgreSQL and pgvector enables semantic search.
  • Extract text from images, PDFs, and documents, then summarize and categorize it.
  • Angular / Ionic frontend, NestJS core API, and FastAPI AI services with OpenAI.
02 / THE PERSON BEHIND THE CODE

Curious by nature.
Builder by instinct.

I’m Nour, a Mathematics & Computing Science graduate based in Vancouver. I enjoy the space where rigorous thinking meets creative problem-solving.

My work spans the full stack: interfaces people interact with, the systems behind them, and AI that makes those systems more useful. I’m drawn to real problems — and to taking an idea all the way to a working product.

Full-stack engineeringMulti-agent systemsSemantic searchApplied machine learningCloud deployment
03 / GET IN TOUCH

Built to make a difference.
Let’s talk.

Have a project, a role, or an idea worth building? Send me a message.