Successfully performed data crawling of the Crunchbase website, followed by data organization and effective circumvention of bot detection methods.

- Turkcell AutoML Platform, Team LeadWe designed a platform where users can upload their data and train different models. I took part in allprocesses, from the design of this platform to its first sale. The company earned 45k€ from the initialrevenue.- Smart SMSA system was created that correctsSMS typos and automatically categorizes them.This system generated a profit of 12.5 million TL.- AI-Based Video InterviewIn the Turkcell unlimited talent program, where morethan 60000 candidates applied and 10010 candidates were selected, and all candidates application documents and interview videos were evaluated using Artificial Intelligence.https://kariyerim.turkcell.com.tr/sinirsiz-yetenek- In the Turkcell Young Talent Program, where morethan 1000 candidates applied and assessed allapplications using our artificial intelligence module.https://kariyerim.turkcell.com.tr/genc-yetenek(Python, Tensorflow, Keras, OpenCV)- TBOT - Chatbot PlatformWe created a chatbot platform in Turkcell. Here, I developed the NLU part and its deployment onOpenShift.https://www.arnavutkoy.bel.tr/(Python, Docker, OpenShift, Tensorflow, Keras,PyTorch)

- Hyper-Personalized Smart ListingRegistered users on the www.beymen.com e-commerce page are provided to see their personalized product rankings in each product category. Measurements achieved up to a 30% increase in product view events. All products on the site are scored for all registered customers. The outputs of the algorithm were also used in mail scenarios.A banner design with a personalized products list was created for each registered user on the www.morhipo.com e-commerce page. All products on the site are scored for all registered customers.Customer Segmentation, Personalization, Cold Start Problem, Product Similarity, Customer Similarity, Python, Airflow- Campaign OptimizationThe machine learning process, which aims to deliver a newly created campaign to the relevant people, has been developed. 66% success was achieved in A / B tests.Campaign Similarity, User Similarity, Customer Targeting, Python, Airflow- Data collectionCreated infrastructure that can handle 90 million events daily. Worked with three different cloud providers.AWS, Azure, GCP, Kubernetes, Python

I worked on a chatbot project as a part-time employee. I developed an algorithm and prepared chatbot infrastructure, services, and chatbot integration. After preparing the infrastructure, I developed an interface for this system.(Java, D4LJ, Weka, Text Similarity, Text Classification, NLP, Spring Boot, Angular JS)

I worked on text files that contained customer complaints. In customer complaints, complaints are often experienced and resolved by another user. I tried to determine the resolution of the new customer complaint by using our written customer complaints and solution data.(Java, Machine Learning, Text Classification, Text vectorization, NLP)I participated in a hackathon at Softtech. In this competition, we made a machine learning application with our team. For a particular contest, I wrote a program using the OpenCV library to take pictures at regular intervals and share them on Instagram.(C++, OpenCV, Python, Instagram API)

I worked on the embedded system project. I have contributed to developing the structure that will provide the necessary communication to see the hardware operations on the web page. I created an API and wrote the document so that others could use API.(NodeJs, C, JavaScript, HTML, CSS, Tessel Board)