Initializing Core Systems
AI & Data Engineer
İstanbul, Turkey

Şehmus Yakut

Architecting AI Systems, Knowledge Graphs & Data Pipelines

I design production-ready Machine Learning models, automated ETL pipelines, and Graph-based RAG architectures. Experienced in computer vision for edge safety devices and medical graph databases.

Computer Engineering @ Yıldız Technical University
2 Academic Publications (MDPI & Springer)

Profile Overview

Current Position

Data Engineer @ Doctor Follow

Building medical ETL pipelines & Neo4j Knowledge Graphs

Education

Yıldız Technical University

B.S. in Computer Engineering (2020 - Jan 2026)

Grant Project

TÜBİTAK 2209-A

Motorcycle Blind Spot Detection

Research

2 Papers Published

MDPI & Springer ICCSA

Core Tech Stack
PythonPyTorchNeo4jETL PipelinesRAG SystemsFastAPIDockerSQL

Featured Engineering Work

Engineering Projects

Systems and architectures I have designed, built, and documented.

Visit GitHub Profile
AI & Machine Learning

TÜBİTAK 2209-A Blind Spot Detection & Warning System

Real-time computer vision and object detection warning system for motorcycle rider safety, funded by TÜBİTAK.

Engineered Solution: Led a 5-member team in designing and implementing a cost-effective real-time blind spot detection and warning system. We deployed computer vision algorithms on edge camera feeds to detect incoming vehicles and trigger timely notifications.

Funding Program

TÜBİTAK 2209-A

Core Technology

Edge Vision AI

PythonOpenCVYOLOPyTorch
AI & Machine Learning

Stress & Fatigue Detection via Facial Expression Recognition

Deep learning facial analysis using ViT and BEiT transformers to identify stress metrics. Official implementation of the ICCSA 2025 paper.

Engineered Solution: Developed an end-to-end deep learning system leveraging transformer-based architectures (ViT, BEiT) trained on the FERPlus dataset, improving model accuracy and establishing robust facial fatigue diagnostics.

Model Architecture

ViT & BEiT

Benchmark Dataset

FERPlus Dataset

PyTorchVision TransformersBEiTOpenCV
AI & Machine Learning

Alzheimer's Disease Classification using Deep Learning

Multi-class MRI scan neural classifier utilizing neuroimaging data to identify Cognitive Normal, MCI, and AD categories. Graduation project implementation.

Engineered Solution: Designed and implemented a multi-class deep learning classifier to categorize MRI brain scans into Cognitive Normal (CN), Mild Cognitive Impairment (MCI), and Alzheimer's Disease (AD) using advanced neural network architectures.

Data Modality

Neuroimaging MRI

Diagnostic Classes

CN vs MCI vs AD

TensorFlowKerasPythonNibabel
Data Platforms & Search Architecture

Travel Assistance: Multi-Language Smart Guide Application

A comprehensive travel assistance web application utilizing Next.js 15 and TypeScript, integrating Google Maps API to provide location-based services, travel tools, and safety alerts with multi-language support (English, Turkish, Kurdish).

Engineered Solution: Engineered a comprehensive, full-stack travel assistant application in Next.js 15 and TypeScript, integrating Google Maps API, multi-currency converter utilities, and real-time translators.

Framework

Next.js 15 & TypeScript

Architecture

MVVM Pattern

Next.js 15TypeScriptTailwind CSSGoogle Maps API
Mobile Development

MedAI: AI-Powered Clinical Assistant & Medicine Planner

Intelligent medicine management and clinical assistant application featuring study planners, OCR prescription scanner, and environment-configured persistent AI chat.

Engineered Solution: Engineered an intent-driven smart hub using Flutter. Integrated Google ML Kit for on-device OCR, built a persistent SQLite database for chat histories, and established a secure environment configuration to route queries to the Gemini API.

Target Platform

iOS & Android

AI Engine

Gemini Model

FlutterDartGoogle ML Kit (OCR)Gemini API
Mobile Development

PharmAI: Intelligent Pharmacology Assistant

A Flutter mobile application designed to serve as a smart assistant in pharmaceutical study and reference environments.

Engineered Solution: Developed a lightweight, offline-first Flutter application framework to store and retrieve key drug formulas, dosages, and interactions rapidly.

Target Platform

Cross-Platform

Core Framework

Flutter SDK

FlutterDart

Career Timeline

Professional Experience

View full history

Data Engineer

Doctor Follow

Istanbul, Turkey02.2026 - 05.2026

  • Develop and orchestrate automated ETL pipelines using Python to structure large-scale medical datasets for improved accessibility.
  • Optimize medical knowledge graphs (Neo4j) to enhance the retrieval speed and accuracy of RAG-based search applications.
  • Design structured data models to facilitate high-speed querying across complex, interconnected medical content nodes.
PythonNeo4j (GraphDB)ETL PipelinesData ModelingRAG SystemsSQL

Network Technologies Intern

Turkcell

Istanbul, Turkey02.2025 - 04.2025

  • Automated routine network administration tasks using custom Python scripts to improve structural efficiency.
  • Monitored and supported core network deployments, verifying topology parameters.
  • Implemented test automation scripts and collaborated on network topology design modifications.
  • Enhanced troubleshooting speed and improved overall network reliability.
PythonNetwork AutomationTopology MonitoringTroubleshootingScripting

Mobile Developer (part-time)

Spor Sepeti

Istanbul, Turkey03.2024 - 10.2024

  • Co-developed the cross-platform SPORSEPETI.app mobile application, focusing on real-time data synchronization.
  • Implemented robust backend data schemas, API integration gates, and dynamic RESTful controllers in Dart.
  • Managed live stream feeds and state containers, ensuring high performance and responsive client views.
DartFlutterRESTful APIsData StreamsBackend IntegrationGit

Technical Competency

Tools & Technologies

Languages & Databases

Python (Pandas, NumPy)Expert
SQLAdvanced
DartAdvanced
CIntermediate
Neo4j (Graph DB)Expert
PostgreSQLAdvanced

Data Engineering & Ops

ETL PipelinesExpert
Data ModelingAdvanced
DockerAdvanced
Data PreprocessingExpert
GitAdvanced

AI & Machine Learning

RAG SystemsExpert
NERAdvanced
LLM/SLM FrameworksExpert
PyTorchAdvanced
TensorFlowAdvanced
scikit-learnAdvanced

Tools & Analysis

MLflowAdvanced
Statistical ModelingAdvanced
Experiment TrackingAdvanced
Data VisualizationAdvanced

Infrastructure Contributions

Real-time GitHub Data
 

Peer-Reviewed Research

Publications

View all papers
MDPIComputers 2026, 15, 77. MDPI

A Multimodal Transformer-Based Framework for Emotion Analysis in Multilingual Video Content

Analyzing emotions in multilingual video content is challenging due to acoustic variations, linguistic expressions, and diverse visual cues. We introduce a multimodal transformer framework designed to fuse acoustic embeddings, visual features, and textual semantic cues, yielding robust emotion classification profiles across multiple language structures.

Authors: Yakut, S., Tuten, Y. T., Caglar, E., Aktas, M. S.

Source Link
Springer, ChamComputational Science and Its Applications – ICCSA 2025 Workshops (Lecture Notes in Computer Science, Vol. 15886)

Facial Stress and Fatigue Recognition via Emotion Weighting: A Deep Learning Approach

This research proposes a deep learning model for detecting facial stress and fatigue metrics. By integrating emotion weighting coefficients on visual expressions, we improve the identification threshold of fatigued states under variable industrial environments, outperforming standard baseline classification pipelines.

Authors: Oskooei, A. R., Caglar, E., Yakut, S., Tuten, Y. T., Aktas, M. S.

Source Link

Supervisor Recommendations

Professional References

"Supervised and mentored Şehmus during his internship at Turkcell, where he contributed to network automation, troubleshooting, and performance monitoring tasks."

İbrahim Karpuz

Senior Network & Security Engineer, Turkcell

karpuz.ibrahim@turkcell.com.tr

"Worked closely with Şehmus during his internship and subsequent freelance collaboration at SPORSEPETI, where we co-developed mobile application backend infrastructure, feature integration, and performance optimization."

Ahmad Khawatmi

iOS Engineer (HaCon) | Co-Founder, SPORSEPETI.app

ak93x.work@gmail.com

Let's Discuss AI & Data Engineering

Based in İstanbul, Turkey. Available for machine learning engineering, data pipeline architecture, and GenAI RAG system development.