Mert Kanıbir



I start with the person who has to use what I build: their constraints, their context, and what should become easier for them. That shapes the whole system, from the model and backend to the smallest interaction. For me, human-centered design is an engineering decision: it determines what I build, what I measure, and whether the result is useful beyond a demo.
Technical focus
AI / Machine learning
Building AI into usable products: LLM integration, adaptive learning logic, agent workflows, and model experimentation. Previous AI architecture and development work includes Learning Ladder AI.
- Python
- Gemini
- Vertex AI
- TensorFlow
- PyTorch
- OpenCV
- OpenPose
Human-centered computing
Designing around real user constraints in accessibility and education. As Co-Founder & CEO of NAVINDOOR, explored Bluetooth beacon positioning and audio-guided indoor navigation through research and prototyping.
- YANA
- Learning Ladder AI
- NAVINDOOR
- Education technology
Full-stack / Systems
Connecting backend logic, web and native interfaces, APIs, databases, and containerized development environments into complete applications.
- Python
- TypeScript
- JavaScript
- React
- Swift
- SwiftUI
- Flask
- Docker
- Supabase
- SQL
Selected tools
Languages
- Python
- TypeScript
- JavaScript
- Swift
- Java
- C/C++
- SQL
- x86-64 Assembly
Interfaces & applications
- Next.js
- React
- SwiftUI
- Tailwind CSS
- Qt Widgets
- WidgetKit
- Canvas
AI & computer vision
- Gemini
- Vertex AI
- LangGraph
- TensorFlow
- PyTorch
- OpenCV
- OpenPose
- ElevenLabs
- Apple Foundation Models
- Flair NLP
- Wan 2.2
Backend & data
- FastAPI
- Flask
- REST APIs
- WebSockets
- Supabase
- MongoDB
- Auth0
- UserDefaults
Cloud & development
- AWS
- Docker
- Git
- GitHub
- CMake
- Make
- GDB
Research & prototyping
- Matplotlib
- Bluetooth LE
- Spatial audio
Research focus
YANA · Accessible AI research
Through YANA (Your ASL Native Assistant), I explored how computer vision and generative AI could support American Sign Language interaction. My completed undergraduate research connected CNN-based gesture recognition, OpenPose skeletal tracking, pose estimation, and generative video in a functional prototype.
- 01 · CaptureSkeletal tracking
- 02 · InterpretGesture recognition
- 03 · GenerateGenerative video
- Semesters of undergraduate research
- 3 semesters
- Peer-reviewed papers synthesized
- 50+
Education
B.S. Computer Science
![]()
Thomas J. Watson College of Engineering and Applied Science
- Watson College Scholars Program
- First-Year Research Immersion
- Computers, Robotics & Engineering Community
- Educational Opportunity Program
Certifications & continued learning
Generative AI Fundamentals of AWS
QA North America
AI for Everybody
QA North America
Essentials of AI for Business
MinnaLearn