# Erdoğan Yasin Peker — full profile

> AI Engineer based in Kırşehir, Türkiye, and a Computer Engineering graduate of Harran University (B.Sc., May 2022 – June 2026; Erasmus+ exchange at AGH University of Krakow, Poland). Currently an Artificial Intelligence Engineer (internship) at Car Studio AI in Kocaeli, Türkiye. Specializes in multimodal AI: deep learning, computer vision, large language models, and audio/signal processing. Built 20+ projects, took 2nd place in Türkiye at the TEKNOOCAK Technology and Innovation Festival, is a 2x Teknofest finalist and team leader in the Healthcare AI category, and holds 13 certifications in ML/AI/data. Open to remote, hybrid, and on-site work.
>
> This is the full-content version of https://erdoganpeker.com/llms.txt — it contains the complete Medora case study, the open-source project catalogue, the FAQ, and the article list.

## Links
- Website: https://erdoganpeker.com (canonical URLs are locale-prefixed: /en/ and /tr/)
- About: https://erdoganpeker.com/en/about/ (TR: /tr/about/)
- Projects: https://erdoganpeker.com/en/projects/ (TR: /tr/projects/)
- Medora case study: https://erdoganpeker.com/en/projects/medora/ (TR: /tr/projects/medora/)
- Blog: https://erdoganpeker.com/en/blog/ (TR: /tr/blog/)
- Blog feed (EN): https://erdoganpeker.com/feed-en.xml
- Contact: https://erdoganpeker.com/en/contact/
- Email: erdoganpeker4@gmail.com
- Phone: +90 553 495 1433
- WhatsApp: https://wa.me/905534951433
- GitHub: https://github.com/ErdoganPeker
- LinkedIn: https://www.linkedin.com/in/erdogan-yasin-peker-b107ba24b/
- Kaggle: https://www.kaggle.com/erdoanpeker
- Medium: https://medium.com/@erdoganpeker

## Machine-readable versions
- English, summary: https://erdoganpeker.com/llms.txt
- English, full content (this file): https://erdoganpeker.com/llms-full.txt
- Türkçe sürüm: https://erdoganpeker.com/llms-tr.txt
- Türkçe, tam içerik: https://erdoganpeker.com/llms-full-tr.txt

## About
Computer Engineer specializing in artificial intelligence and machine learning; graduated from Harran University in June 2026. Lives in Kırşehir, Türkiye, and is open to remote, hybrid, and on-site work. Currently works as an Artificial Intelligence Engineer (internship) at Car Studio AI in Kocaeli, on computer vision and deep learning projects. Previously an AI Engineer at CerebrAI-VorTX (January 2025 – May 2026), working on computer vision and data processing projects. During internships at Bluesense AI and eCloud Software Technologies, developed YOLO-based computer vision models, dermatological analysis systems, and a medical-domain LLM. Reached the Türkiye finals in two different Teknofest categories and led a team in the Healthcare AI category. Shares machine learning projects through Kaggle and GitHub.

## Experience
Reverse-chronological; one role per block.

### Artificial Intelligence Engineer — Car Studio AI (CURRENT ROLE)
July 2026 – present · Kocaeli, Türkiye · Internship (On-site)
- Development on AI projects focused on computer vision and deep learning.
- Preparing image datasets and running the model training pipeline.

### AI Engineer — CerebrAI-VorTX
January 2025 – May 2026 · Şanlıurfa, Türkiye · Full-time
- Active role in design and development of AI projects focused on computer vision and data processing.
- End-to-end model pipeline management: data collection, preprocessing, training, evaluation, and deployment.
- Optimization of deep learning models and hyperparameter tuning.
- Producing technical documentation and model reports.

### AI Engineer Intern — Bluesense AI
June 2025 – August 2025 · Istanbul, Türkiye · Internship (On-site)
- Developed and optimized YOLO-based computer vision models for the Smart Beauty platform.
- Active role in an international project developing real-time mobile dermatological analysis systems.
- Deep learning models with EfficientNetV2, MobileNetV3, Swin-ViT, and MobileViT architectures.
- Improved classification accuracy using hard voting and soft voting ensemble methods.
- Cleaned, annotated, and preprocessed large-scale image datasets.
- Built end-to-end ML pipelines to enable real-time inference.
- Participated in scientific paper preparation for IEEE/Elsevier journals.
- Completed the internship with a "Performance-Based Excellence" certificate.

### AI Engineer Intern — eCloud Software Technologies
June 2025 – August 2025 · Şanlıurfa, Türkiye · Internship (Remote)
- Developed a medical-domain LLM capable of understanding clinical texts.
- Fine-tuned open-source LLMs on medical terminology, diagnoses, and treatment guidelines.
- Integrated the LLM into a web platform via RESTful API.

### Trainee — Game and Application Academy
November 2024 – August 2025 · Istanbul (Remote) · Full-time
- Practical training in AI and data processing.
- Projects in ML, deep learning, and computer vision.

### Team Leader — Teknofest, Healthcare AI
May 2024 – September 2024 · Şanlıurfa · Competition
- Led a team developing AI & computer vision solutions for cancer detection.
- Built a deep learning model for cancer detection in medical images.

Note: job locations above are historical, per role. Current place of residence is Kırşehir, Türkiye.

## Education
- B.Sc. Computer Engineering — Harran University (May 2022 – June 2026, Şanlıurfa, Türkiye) — graduated
- Computer Engineering (Erasmus+ Exchange) — AGH University of Krakow (October 2025 – March 2026, Kraków, Poland). Courses: artificial intelligence, computer vision, systems programming, deep learning.

## Skills
- AI & Machine Learning: PyTorch, TensorFlow, Scikit-learn, Whisper, EfficientNetV2, MobileNetV3, Swin-ViT, MobileViT
- Programming Languages: Python, C, C++, Java, TypeScript
- Libraries & Tools: Flask, FastAPI, Next.js, Docker, Git, Roboflow, RESTful API
- Specializations: Machine Learning, Deep Learning, Reinforcement Learning, Computer Vision, NLP, Audio AI, LLM Fine-tuning, CNN, Transformer, Signal Processing, Multimodal AI

## Certifications
- Google Advanced Data Analytics — Coursera (June 2025)
- Google Project Management — Coursera (May 2025)
- Data Science & AI — Google Academy (May 2025)
- Web App Development — Google Academy (May 2025)
- Entrepreneurship — Google Academy (May 2025)
- Feature Engineering + Data Preprocessing + Mass Detection in Mammography — Presidency of Türkiye, Digital Transformation Office (May–June 2024)
- Image Processing — Udemy (August 2024)
- General AI — Udemy (May 2024)
- Deep Learning — Udemy (August 2023)
- Machine Learning — Udemy (July 2023)
- Data Visualization — Udemy (June 2023)
- Data Science — Udemy (June 2023)
- Python Programming — Udemy (May 2023)

## Languages
- Turkish — Native
- English — B2+ (Advanced)

## Notable achievements
- Türkiye 2nd place at the TEKNOOCAK Technology and Innovation Festival (2026), in the "Afet, Güvenlik, Dayanıklılık ve Eğitim Teknolojileri" (Disaster, Security, Resilience and Education Technologies) category, organised by the Ülkü Ocakları Education and Culture Foundation
- Selected for the Yapay Zeka ve Teknoloji Akademisi (AI and Technology Academy) as one of 2,000 main scholarship recipients out of roughly 25,000 applicants — a programme run by Google, Girişimcilik Vakfı, and T3 Vakfı with the support of the Ministry of Industry and Technology and the Presidency's Digital Transformation Office; programme completed
- Completed the Bluesense AI internship with a "Performance-Based Excellence" certificate
- Built Medora, a production-grade multimodal health AI platform (201 API routes, 10 ML models, web + Flutter mobile)
- 15+ additional confidential production projects under NDA (healthcare, automotive, food, agritech), shown anonymously
- 2x Teknofest finalist; team leader in the Healthcare AI category (2024)
- 100th place in the BTK Academy Datathon with a student-performance evaluation model
- 20+ open-source projects in deep learning, computer vision, NLP, and audio AI
- 13 certifications (Google, Coursera, Udemy, Presidency of Türkiye DTO) in ML/AI/data
- Participated in the PROJEXeleration 2026 acceleration programme (participation only — not an award, a placement, or a selection)

---

## Flagship project: Medora — full case study
Source: https://erdoganpeker.com/en/projects/medora/ · CONFIDENTIAL PROJECT — CASE STUDY

Tagline: "AI for your health."

A multimodal health AI platform that unifies image, audio, signal, and text processing under one roof. Designed and built end-to-end on FastAPI, React, Flutter, and PostgreSQL — running today on Docker.

### Medora in numbers
- 201 — API routes
- 10 — Disease prediction models
- 30+ — Mobile screens (pixel-mirror)
- 3300+ — Vital measurements processed
- 5/5 — Containers healthy
- WCAG 2.1 AA — Accessibility

### Problem & Solution
Health data is not one thing: a prescription photo, a skin lesion image, a heart-rate series, and a patient's complaint in free text all belong to the same person, yet each demands a completely different processing technique. Most systems pick one and leave the patient with a fragmented experience.

Medora unifies these modalities in a single platform: computer vision reads the prescription and the lesion, Whisper transcribes speech, a signal-processing layer catches anomalies in vital measurements, and a health-focused Turkish voice agent translates all of it into language the patient understands — with 112 emergency routing and a strict "not medical advice" safety frame. Patient, doctor, clinic manager, and admin each get their own role-based interface, and web and mobile deliver a pixel-identical experience.

### Competency Matrix — what exactly was built in Medora?
This project is evidence of end-to-end engineering breadth, not a single specialty.

**Computer Vision / Deep Learning**
- Skin lesion analysis: lesion type, malignancy risk, class scores, and a see-a-doctor recommendation (gpt-4o vision)
- X-ray image analysis
- Prescription OCR: extracting drug name + dose + usage from real prescription images at 98–99% confidence, mapped to a drug database and pushed into e-prescriptions.

**Audio Processing**
- Speech recognition with Whisper
- spoken responses via tts-1 with auto-playback on every turn (voice-first UI)
- real-time audio streaming over WebSocket
- microphone-permission fallbacks.

**Signal Processing**
- Heart rate, blood pressure (systolic/diastolic merging), temperature, SpO₂, and glucose streams
- anomaly detection and automatic alert generation over 3,300+ measurements (vital_anomaly service).

**Machine Learning**
- 10 disease prediction models (GradientBoosting / sklearn pipelines)
- the heart model trained on real data
- a dedicated Flask ML serving layer.

**LLM & Agentic AI**
- A health-focused Turkish voice agent (gpt-4o-mini)
- 112 emergency routing and a "not medical advice" safety disclaimer
- RAG with semantic search on pgvector
- agentic report generation
- a configurable LLM layer (local/remote models + A/B testing).

**Backend & API**
- 201 routes on FastAPI
- async SQLAlchemy
- real-time WebSocket channels
- rate limiting
- nginx reverse proxy.

**Database**
- PostgreSQL + pgvector
- Redis
- a 17-module idempotent demo seed system
- a realistic dataset of 22 active patients, 8 doctors, and thousands of clinical records.

**Web & Mobile**
- React/Vite web (~85 kB main bundle, code-split)
- Flutter mobile: 30+ screens, pixel-mirrored against the web app
- flutter analyze: 0 errors / 0 warnings
- 14 custom web UI components + 10 Medora Flutter widgets.

**Security & Identity**
- JWT + argon2/bcrypt
- MFA
- role-based access (patient / doctor / clinic manager / admin)
- WebSocket auth (Sec-WebSocket-Protocol: bearer.<jwt>)
- rate limiting.

**Healthcare Integrations**
- FHIR
- e-Nabız
- e-Prescription
- e-Report
- MHRS
- an e-Devlet flow
- billing + e-Invoice
- PDF generation with full Turkish character support (embedded DejaVu fonts)
- Google Meet video consultations
- family profiles
- medication reminders with adherence scoring
- a moderated community module.

**Accessibility & Design System**
- The "Modern Medical" design system (Medical Teal #0EA5A6, Inter)
- WCAG 2.1 AA: modal focus traps + ESC, arrow-key tab navigation, skip-to-content link, aria-labels
- measured contrast fixes (6.09:1)
- page transitions and KPI stagger animations.

### Architecture
Medora is a containerized system of five cooperating services: the FastAPI backend (201 routes), the React/Vite web client, a Flask ML server, PostgreSQL (with pgvector), and Redis — fronted by nginx. The Flutter mobile app consumes the same API. Three separate Docker Compose profiles (local / dev / prod) and multi-stage builds carry the same codebase from development to production; all five containers run healthy.

- Clients: React / Vite web · Flutter mobile
- Edge: nginx reverse proxy
- Services: FastAPI backend — 201 routes · Flask ML server
- Data layer: PostgreSQL + pgvector · Redis
- Docker Compose ×3 profiles (local / dev / prod) · multi-stage build · 5/5 containers healthy

### Technology Stack
- AI/ML: gpt-4o vision, gpt-4o-mini, Whisper, tts-1, scikit-learn, GradientBoosting, RAG, pgvector
- Backend: FastAPI, Flask, SQLAlchemy (async), WebSocket, Redis, nginx
- Frontend & Mobile: React, Vite, Flutter, Inter, Tailwind-based design system
- Data & Infrastructure: PostgreSQL, pgvector, Docker Compose ×3, multi-stage build
- Quality: Playwright E2E, flutter analyze, WCAG 2.1 AA

### Engineering Quality
- Playwright E2E tests; a clean baseline across 4 roles × ~60 routes.
- A stable release after multiple bug-hunt and fix cycles.
- flutter analyze: 0 errors / 0 warnings.
- WCAG 2.1 AA verified by measurement (including 6.09:1 contrast fixes).
- A 17-module idempotent seed: the system boots with realistic data in any environment with a single command.
- 3 environment profiles (local/dev/prod), multi-stage Docker builds, 5/5 containers healthy.

### Confidentiality Note
Medora's source code is closed for commercial and confidentiality reasons. A live demo, an architecture walkthrough, and a code review can be arranged upon request. Medora is just one of 15+ projects under NDA.

---

## Open-source projects
Public repositories on GitHub. Detail pages live under https://erdoganpeker.com/en/projects/.

- Student Evaluation — BTK Datathon — Student-performance evaluation that placed 100th in the BTK Academy Datathon; ML/DL models with comprehensive exploratory data analysis. (https://github.com/ErdoganPeker/Student-Evaluation-System-BTK-Academy-Competition)
- India House Price Prediction — House-price prediction reaching an R² of 0.93–0.94 with an MSE of 2×10⁻⁶ on scaled targets; ML/DL model comparison and exploratory data analysis. (https://github.com/ErdoganPeker/India-House-Price-Prediction)
- Obesity Level Prediction — Predicting obesity status from behavioral and physical features with ML/DL, comprehensive EDA, and feature engineering. (https://github.com/ErdoganPeker/Obesity-Level-Prediction)
- Titanic Survival Prediction — Predicting Titanic passenger survival with ML/DL, exploratory data analysis, and data visualization on the classic dataset. (https://github.com/ErdoganPeker/Titanic-Survival-Prediction)
- Store Sales Analysis — Multi-store sales trend analysis and time-series forecasting with an EDA- and visualization-heavy ML/DL workflow. (https://github.com/ErdoganPeker/Store-Sales-Evaluation)
- Smart Cleaning Robot — A C++ cleaning-robot simulation that reads obstacle maps and optimizes cleaning paths with BFS/DFS-based path planning. (https://github.com/ErdoganPeker/Smart-Cleaning-Robot)
- Bus Management System — A bus reservation and customer management system in Java and SQL with full CRUD operations and multi-route schedule management. (https://github.com/ErdoganPeker/Bus-Management-System)
- Disaster Relief System — A disaster-relief system in C managing victim records with linked lists and a priority queue, using binary search for fast lookups. (https://github.com/ErdoganPeker/Disaster-Relief-System)
- Pet Care Simulator — A C simulation game where you keep a virtual pet healthy and happy, built on a state machine with save/load via file I/O. (https://github.com/ErdoganPeker/Pet_Care_Simulator)
- Car Price Prediction (Turkish Market) — Car-price prediction on Turkish used-car market data comparing linear regression, random forest, XGBoost, and neural networks, with full EDA. (https://github.com/ErdoganPeker/Araba-Fiyat-Tahmini)
- Student Performance Evaluation — Predicting academic performance from behavioral data with ML/DL; comparative analysis of 8+ algorithms with EDA and visualization. (https://github.com/ErdoganPeker/Student-Performance-Evaluation)
- Weight Tracker — FastAPI + React + Expo — Calendar-based weight and calorie tracking: FastAPI backend, React PWA, Expo mobile, and Docker, with Gemini AI food recognition and a 607-item Turkish food database. (https://github.com/ErdoganPeker/weight-tracker)
- Real-Time Voice & Emotion Processing — Real-time speech-to-text with Whisper and emotion recognition from audio, streamed over Flask + WebSocket with live microphone capture. (https://github.com/ErdoganPeker/Real-Time-Voice-and-Emotion-Processing)
- Reinforcement Learning Projects — Sub-projects implementing Q-Learning, Deep Q-Learning, and Convolutional Q-Learning from scratch with PyTorch and OpenAI Gym. (https://github.com/ErdoganPeker/Reinforcement-Learning-Projects-Q-Learning-Deep-Q-Learning-and-Convolutional-Q-Learning.)

---

## FAQ

**Who is Erdoğan Yasin Peker?**
Erdoğan Yasin Peker is an AI engineer based in Kırşehir, Türkiye. He graduated in Computer Engineering from Harran University in June 2026, completed an Erasmus+ exchange at AGH University of Krakow in Poland, and has built more than 20 projects in computer vision, large language models, audio, and signal processing. He currently works as an Artificial Intelligence Engineer (internship) at Car Studio AI in Kocaeli, Türkiye.

**What does Erdoğan Yasin Peker specialize in?**
His specialty is multimodal AI: computer vision (YOLO, CNN, and Vision Transformer architectures), LLM fine-tuning, audio processing with Whisper, and vital-signal analysis. His daily toolkit is PyTorch, TensorFlow, scikit-learn, FastAPI, and Docker; he develops in Python, C, C++, Java, and TypeScript.

**What is Medora?**
Medora is a confidential (NDA) multimodal health AI platform that Erdoğan Yasin Peker designed and built end-to-end: 201 FastAPI routes, 10 disease-prediction ML models, a React web client, and a Flutter mobile app with 30+ screens. It covers skin-lesion analysis, prescription OCR, speech recognition with Whisper, anomaly detection over vital measurements, and WCAG 2.1 AA accessibility. The source is closed; a live demo is available on request.

**Where has Erdoğan Yasin Peker worked?**
Since July 2026 he has been an Artificial Intelligence Engineer (internship) at Car Studio AI in Kocaeli, Türkiye — this is his current role. Before that he was an AI Engineer at CerebrAI-VorTX from January 2025 to May 2026. He also completed AI engineering internships at Bluesense AI (YOLO-based computer vision for the Smart Beauty platform and dermatological analysis systems; finished with a "Performance-Based Excellence" certificate) and eCloud Software Technologies (a medical-domain LLM), finished the Game and Application Academy trainee program, and led a team in the Teknofest Healthcare AI category.

**What certifications does he hold?**
He holds 13 certifications: Google Advanced Data Analytics and Google Project Management (Coursera), Google Academy programs in data science, web development, and entrepreneurship, a Presidency of Türkiye Digital Transformation Office program on mass detection in mammography, and Udemy courses in machine learning, deep learning, image processing, and data science.

**Is he open to job offers and collaborations?**
Yes. He is based in Kırşehir, Türkiye and open to remote, hybrid, and on-site work. He speaks Turkish (native) and English (B2+). He can be reached at erdoganpeker4@gmail.com or on WhatsApp (+90 553 495 14 33), and usually replies the same day.

**What are his competition achievements?**
He took 2nd place in Türkiye at the TEKNOOCAK Technology and Innovation Festival (2026), in the "Afet, Güvenlik, Dayanıklılık ve Eğitim Teknolojileri" (Disaster, Security, Resilience and Education Technologies) category organised by the Ülkü Ocakları Education and Culture Foundation. He reached the Türkiye finals in two different Teknofest categories and led a team in the 2024 Healthcare AI category working on cancer detection in medical images. He also placed 100th in the BTK Academy Datathon with a student-performance evaluation model. He was selected as one of 2,000 main scholarship recipients out of roughly 25,000 applicants for the Yapay Zeka ve Teknoloji Akademisi (Google, Girişimcilik Vakfı, and T3 Vakfı, supported by the Ministry of Industry and Technology and the Presidency's Digital Transformation Office) and completed the programme. Separately, he participated in the PROJEXeleration 2026 acceleration programme — participation only, not a placement or an award.

---

## Articles

### Building a Multi-Platform Calorie Tracker: FastAPI + Docker + Expo
Published 2026-01-15 · https://erdoganpeker.com/en/blog/weight-tracker-fastapi-expo/
How I built a weight and calorie tracking app with Gemini AI food recognition, 607 Turkish foods, Docker single-command deployment, and a cross-platform Expo mobile app that works offline.

### Vehicle Damage Detection with YOLOv8-seg: A Multimodal Web + Desktop + Mobile System
Published 2025-10-20 · https://erdoganpeker.com/en/blog/yolo-vehicle-damage-detection/
How to build a full multimodal system for vehicle damage detection using segmentation models, with clients for web, desktop, and mobile. The story of a Next.js + Tauri + React Native pipeline.

### Real-Time Speech & Emotion Analysis: What I Built with Whisper + NLP
Published 2025-06-15 · https://erdoganpeker.com/en/blog/ai-speech-emotion-analysis/
How to build a system that transcribes microphone audio in real-time and performs emotion analysis? I share how I built this using Whisper, Flask, and WebSockets.

Last updated: 2026-08-02