AI / ML
- Python
- TensorFlow
- Keras
- scikit-learn
- OpenCV
COMPUTER ENGINEER
TÜRKİYEAI · AUTOMATION
SOFTWARE SYSTEMS
ARTIFICIAL INTELLIGENCE · SOFTWARE · AUTOMATION
Adil Buğra Aytar.
I build AI applications, automation systems and practical software.
01About
Beyond notebooks, working systems.
- Education
- Computer Engineering · SDU
- Focus
- AI · Automation · Software
- Approach
- Build · Test · Improve
I am a Computer Engineering graduate of Süleyman Demirel University. I work with machine learning, deep learning and large language models, building Python software that spans data preparation, model development, RAG assistants and computer vision.
I gained experience in AI and data science development at Zen Tech and Verkosis, and in LLM evaluation at Outlier AI. I care about usability as much as model accuracy: well-prepared data, measurable tests and an application that fits the problem.
02Technologies
Tools chosen for the problem.
Automation
- n8n
- API integrations
- Workflow automation
Backend / Infrastructure
- Python
- SQL
- Docker
LLM / AI
- LLMs
- RAG
- OpenAI API
- Azure OpenAI
- Ollama
- LLaMA
Web
- HTML
- CSS
- JavaScript
03Projects
Applied AI, automation and software systems.
A collection of working projects built around data, real constraints and practical workflows.
LLM System
SDU Regulations RAG Assistant
A RAG assistant that answers questions about university regulations using relevant document context. I designed the data preparation, embedding and query workflow end to end.
Explore project
Problem: Make university regulations accessible through natural-language questions. My contribution: Data collection and cleaning, embedding generation, document retrieval and the answer-generation workflow. Implementation: Model setup and optimization with LLaMA 3.2 Vision, RAG, Docker, Azure OpenAI and OpenWebUI. Tested with different question scenarios.
Audio processing · Deep learning
Vocalize AI
A deep learning project that classifies urban sounds into 10 categories, using MFCC features extracted from audio recordings.
Explore project
Data: The 10 sound classes in UrbanSound8K. Approach: Extract 40 MFCC features with Librosa and train a dense neural network with TensorFlow / Keras. The repository includes preprocessing, model and training code, plus an analysis notebook.
Data analysis · Machine learning
Customer Segmentation AI
A machine learning project that groups customers by annual income and spending score, visualizing five segments with K-Means.
Explore project
Data: 200 customers from the Mall Customer Segmentation dataset. Approach: Normalize with MinMaxScaler, explore the number of clusters with the elbow method and segment with K-Means. The repository includes data preparation, model and visualization code.
Computer vision · CNN
Object Recognition & Classification
A deep learning project that classifies CIFAR-10 images into 10 categories. It covers image preprocessing, CNN training and visualization of predictions.
Explore project
I developed a CNN image classifier using CIFAR-10. After preprocessing and training, I visualized predictions on test images. The experiment steps are available in the Jupyter notebook.
Time series · LSTM
Gold Price Prediction
A time-series modeling project using historical monthly gold prices. It generates a six-month prediction sequence with an LSTM model.
Explore project
I worked with 58 monthly observations from January 2020 to October 2024. I scaled the data, created 12-month windows and generated a six-month forecast using a two-layer LSTM. Data preparation, training and result charts are available in the repository.
Machine learning · Desktop application
HomePrice AI
A regression project that estimates home prices from features such as room count, floor area and building age. A Python desktop interface makes the model accessible to users.
Explore project
I developed a multiple linear regression model using regional housing data. I designed a Python desktop application that turns user-provided property features into model estimates. The repository includes data and analysis notebooks.
04Experience
Learning by building, testing and shipping.
Outlier AI
AI Trainer / LLM Evaluation Specialist
Remote · Produced, edited and evaluated data for large language models. Checked model outputs for accuracy and consistency, and worked on prompt development and model performance improvements.
Zen Tech Elektronik
AI & Data Science Developer
Isparta · Developed an LLM assistant for university regulations. Worked across data collection and cleaning, model setup, the RAG query workflow and user experience.
Verkosis Bilgi Teknolojileri
AI & Data Science Developer
Isparta · Developed machine learning and deep learning models with TensorFlow and Keras. Focused on improving model performance through preprocessing, feature engineering and hyperparameter optimization.
05Education
The foundations of engineering.
Süleyman Demirel University
Computer Engineering
2019 — 2024
Süleyman Demirel University
Computer Programming
2015 — 2017
06Contact
A direct line for practical work.
Get in touch about a project, a collaboration or a new opportunity.