My Portfolio

A selection of the AI, digital-health, and computer-vision products I’ve designed and built as founder–CTO. For detailed technical docs or demos, feel free to contact me.

1. PureFit Coach — AI-Assisted Personal Training

Role: Co-Founder & CTO / Lead Engineer

Description: Architected and built an AI-assisted online fitness coaching platform end to end, including onboarding, assessments, personalized workout programming, guided workouts, coach dashboards, payments, analytics, and browser-based computer vision.

Engineering: Built movement-analysis workflows for repetition counting and form feedback while balancing browser performance, latency, usability, and infrastructure cost.

Tech Stack: TensorFlow.js, MediaPipe/BlazePose, Node.js, Express, MongoDB, Next.js, React, AWS (EC2, S3, CloudFront)

2. PureFit AI Coaching System

Role: Co-Founder & CTO

Description: PureFit is an AI-powered coaching system used in a physical studio and in a web coach. It leverages 3D pose estimation, form correction, rep counting, range-of-motion analysis, and posture scoring to deliver personalised strength and corrective workouts with minimal on-floor staff.

Engineering: Deployed across two real-world staff-light fitness studios, where workout and movement data was used to improve exercise logic, form-feedback rules, and system reliability.

Tech Stack: TensorFlow.js, MediaPipe/BlazePose, Node.js, Express, MongoDB, Next.js, React, AWS (EC2, S3, CloudFront)

3. Bewell Platform
(Mobile App, Marketplace & Insurance Integration)

Role: Founder & CTO

Description: Bewell is a digital-health platform that combines an AI-assisted mobile app and an online dietitian marketplace to help users build healthy habits in nutrition, exercise, and psychology. The platform includes a special web app integration for Türkiye Sigorta, one of Turkey’s largest insurance companies, giving policyholders access to Bewell content inside the insurer’s app.

Tech Stack: React Native, Expo, Next.js, Node.js, Express, MongoDB, AWS, NLP/ML for content and recommendations

4. My Fitness Trainer – Holistic Wellness App

Role: Co-Founder & CTO / Product Lead

Description: My Fitness Trainer was an early holistic wellness app in Turkey that combined exercise, nutrition, and psychology coaching and later evolved into the Bewell ecosystem. I led the product and managed the development team.

Impact: Selected into İTÜ Çekirdek incubation, achieved strong early user growth and retention.

Tech Stack: React Native, Node.js, MongoDB, AWS

5. Zone Fitness — Tech-Enabled Hybrid Training Business

Role: Founder & CTO / Product Lead

Description: Built the digital backbone of a 3-location hybrid-training chain, including scheduling algorithms, workout-logic systems, and automated member workflows.

This became the foundational architecture for my later AI coaching systems.

Tech Stack: React Native, Node.js, MongoDB, AWS

6. Selected Technical Experiments

Custom NLP APIs: Developed internal APIs for Named Entity Recognition, classification, location extraction, and text matching, all without relying on third-party ML services.

Crypto Trading Bot: Built a fully automated cryptocurrency trading bot in Python to research algorithmic trading and risk controls.

AI Tic-Tac-Toe: Implemented a reinforcement-learning-based tic-tac-toe agent using custom ML logic as a playground for RL concepts.

How I Work

I’m most effective in 0→1 environments where requirements are still evolving. I typically work across product definition, architecture, frontend, backend, data models, AI/CV integration, deployment, analytics, and iteration with real users.

I favor pragmatic architectures and simple systems that can ship quickly, generate real usage data, and evolve as requirements become clearer.