Özge Kaçmaz · Computer Engineer

I turn complexity into products that work.

I combine machine learning, knowledge graphs, and reliable software architecture to build decision-support systems and digital products for real-world problems.

Professional portrait of Özge Kaçmaz

How I engineer

  1. Understand
  2. Model
  3. Build
  4. Validate

I turn complex problems into explainable, usable systems.

Selected Work

Six selected projects with engineering decisions documented from model to interface.

IFS Türkiye R&D internship project2026

Predictive Maintenance

A decision flow connecting sensor data to explainable risk and human-approved maintenance actions.

  • AI4I 2020 · approximately 10,000 synthetic records
  • SHAP explanations and saved evaluations
  • Python
  • Django REST
  • React
  • PostgreSQL
  • SHAP
SensorRisk + SHAPPriorityHuman approval
Conceptual flow from sensor signal to explainable risk and human approval. Sensor inputs → ML analysis → Decision context → Human approval.

Bachelor’s thesis2026

ThreatGraph AI

An inspectable graph connecting threat artifacts, tactics, and defense relationships.

  • 651 nodes · 2,567 relationships
  • Inspectable Neo4j graph with Cypher
  • Python
  • FastAPI
  • Neo4j
  • Cypher
  • Random Forest
ArtifactTacticDefenseRelation
Conceptual graph of artifact, tactic, and defense relationships. Threat data → Neo4j graph → Analysis layer → Review interface.

AI and backend project2026

PDF RAG & MCP

A retrieval flow from document chunks to cited answers for web and MCP clients.

  • PDF chunks retain page metadata
  • Local embeddings and ChromaDB retrieval
  • Python
  • FastAPI
  • ChromaDB
  • RAG
  • MCP
PDF chunksRetrievalCited answerMCP client
Flow from document chunks and MCP client to a cited answer. PDF ingestion → Index → Retrieval + LLM → API + MCP.

AI application2026

AI Question Solver

Routes questions through OCR and visual reasoning paths, then evaluates the outcomes.

  • OCR, Vision, Hybrid, and Adaptive paths
  • 140 Pytest tests (repository report)
  • Python
  • FastAPI
  • OpenCV
  • Tesseract
  • Pytest
OCRVisionAdaptive routeEvaluation
OCR and Vision paths merge at adaptive routing and evaluation. Image question → Route selection → Solution → Evaluation.

More Featured Projects

User / AdminAuthorized APIApplication + dataForecast service
Role paths, authorized API, data layer, and external forecast service. React client → ASP.NET Core API → Application layer → Data + forecast.

IFS Türkiye R&D internship project2026

IFS Aurora: Weather Management App

An enterprise-style full-stack application built with Clean Architecture, JWT access/refresh flows, and role-based authorization, combining user administration, managed weather records, and live forecasts.

Role-based administration · live forecast · weather CRUD

  • ASP.NET Core
  • React
  • TypeScript
  • PostgreSQL
TextLanguage checkSummaryPDF / TTS
Flow from text through validation to summary and PDF/TTS output. Text request → Summarization → Output checks → Consumption.

IFS Türkiye R&D internship project2026

AI Summarization Platform

A bilingual LLM platform that turns summarization into a complete product flow with language/output validation, user history, PDF export, text-to-speech, feedback, and controlled access.

TR/EN validation · PDF · TTS · feedback

  • Python
  • Django REST
  • LLM
  • JWT

Earlier & Client Work

JWT Authentication Notes App

A full-stack practice project covering authentication, protected flows, and data lifecycle with React, REST APIs, SQL, and JWT.

OCR-Based Identity Document Extraction

An image preprocessing, Tesseract OCR, field extraction, and database import workflow.

Dent Ege Dental Clinic Website

An SEO-focused web product built with React 19, TypeScript, and Vite to support service discovery, trust, local visibility, contact, and appointment journeys.

GitHub

About Me

I am a Computer Engineering graduate from Pamukkale University. I bring together machine learning, knowledge graphs, document intelligence, and backend systems to build explainable, usable product flows—not just working prototypes. During my IFS Türkiye R&D internship, I developed three end-to-end projects spanning enterprise web development, LLM-based summarization, and predictive-maintenance decision support. I complement my technical work with strong communication, curiosity, and interdisciplinary thinking.

I direct my curiosity not only toward new tools, but toward why a system is dependable and understandable.

Experience

June–August 2025

SQL Developer Intern · Hotech

Remote

  • Stored procedures, queries, and views with Firebird and IBExpert
  • Analytical queries, database documentation, and performance-focused improvements

Skills

I use technologies for the role they serve in a project, not as a checklist.

Programming

Python · SQL · C# · TypeScript · JavaScript · C

AI & ML

scikit-learn · Random Forest · SHAP · RAG · embeddings · knowledge graphs · multimodal LLM · OCR

Backend & applications

FastAPI · Django REST Framework · ASP.NET Core · REST API · JWT · MCP · React · Entity Framework Core

Data

PostgreSQL · Neo4j · Cypher · ChromaDB · Firebird · Power BI

Tools & testing

Docker · Git · Pytest · Playwright · OpenCV · Tesseract · IBExpert

Education & Certifications

2022—2026

Pamukkale University

B.Sc. in Computer Engineering · September 2022–August 2026

  • Inci Foundation Scholarship Recipient · 2023–2026

CERTIFICATIONS

Continued learning

  • Google · 2026
    Foundations: Data, Data, Everywhere
    Ask Questions to Make Data-Driven Decisions
  • Neo4j GraphAcademy · 2025
    Neo4j & Cypher Fundamentals · Graph Data Modeling · Importing Data

References

Zeynep Kıcıkoğlu

Software Engineer · IFS Türkiye

Mentor engineer during the IFS Türkiye R&D internship

Assist. Prof. Fatmana Şentürk

Pamukkale University

Academic advisor for the ThreatGraph AI bachelor’s thesis

Contact

Let’s build something that matters.

Get in touch about opportunities in software development and artificial intelligence.