Hi, I'm Berke
I am experienced with Machine Learning and Data Science.
Experienced with Python, Java, C, C#, SQL, Data Science, Machine Learning and Deep Learning frameworks.
Currently working as Data Scientist at Getir.
About
Hi, I am Berke, I am interested in building machine learning applications. I love to read about Recommender Models, LLMs, and building ML applications.
Also interested in building scalable and low latency real-time systems.
- Languages: Python, Java, C, JavaScript, C#, Bash
- Databases: MySQL, PostgreSQL, MongoDB, Redis, S3
- Libraries: NumPy, Pandas, Seaborn, Apache POI
- Frameworks: scikit, pytorch, Django, fastAPI, React.js
- Tools & Technologies: Git, Docker, AWS, Kubernetes, JIRA
- Other: AI, Algorithms, Optimization Methods, Operating Systems, Stochastic Methods
- Languages(1): English, Turkish, German - Beginner
Experience
- Responsible for end-to-end process of creating and deploying recommender models, working closely with stakeholders to ensure optimal performance and continuous improvement with custom KPIs and A/B testing.
- Implemented matrix factorization model with LightFM to enhance the product details page with more relevant product recommendations.
- Built offline evaluation library for recommender systems that enabled faster iterations of recommendation engines.
- Developed and deployed a job recommender model in production utilizing NLP topic modeling that serves 20000 customers daily accelerated by NVIDIA Rapids for text preprocessing and inference.
- Built a SOTA double machine learning (causal inference) pipeline that delivers highly accurate estimates of price elasticity, allowing for more effective pricing optimization and promotion offers.
- Created Tableau dashboard that is updated by live model for decision making by pricing team.
- Developed optimization model for courier shift assignment resulting in $500.000 annual savings.
- Tools: Python, FastAPI, AWS SageMaker, MongoDB, SQL, RedShift, S3, Airflow, MLflow
- Researching AI supported presentations with Prof. Metin Sezgin.
- Deployed Pytorch model to production for handwriting recognition and classification.
- Improved app fluidity with JavaScript by reducing repetitive API calls.
- Implemented unit tests for full functionality coverage, created CI/CD pipeline.
- Created pipeline for pre-processing drawing data, training and optimizing SVM model and logging results.
- Tools: Python, JavaScript, Pytorch, Scikit-learn, python-pptx
- Created a web application that retrieves and sends authentication codes to our sales force in real-time.
- Implemented employee skills vectorization for clustering in gamification project.
- Created a Java app that reads data stored on Excel documents and transfers them to JSON.
- Transformation and data processing on more than 1200 clients unstructured data.
- Contributed to pre-processing step of ML project for credit risk scoring. Tools: Python, Java Spring, React Js.
- Tools: Java, Python, MongoDB, SQLite, Swing, React
- Created projects using Azure Cognitive Services (Vision, Text analytic, Search API, ML).
- Received mentorship, attended to variety of courses about Azure Services.
- Tools: C#, Python, Azure Services
- Developed a 2D Android platformer game using C# and Unity.
- Tools: C#, Unity
Education
Istanbul, Turkey
Degree: Bachelor of Engineering in Computer Engineering
CGPA: 3.45/4.0
- Intro. to Machine Learning
- Natural Language Processing
- Computer Vision with Deep Learning
- Computer Networks
- Data Privacy & Security
- Operating Systems
- Algorithms and Complexity
- Programming language concepts
- Advanced Programming
Relevant Courseworks:
Double Major
Degree: Bachelor of Engineering in Industrial Engineering
- Optimization Methods
- Stochastic Methods
- Applied Statistics
- Planning and Forecasting
Relevant Courseworks:
Vancouver, Canada
Degree: Computer Science
Exchange Student
Received Study in Canada Scholarships
- Artificial Intelligence
Relevant Courseworks:
Projects
![quiz app](/assets/img/morse.png)
Morse translator on FPGA Board
- FPGA Board translates morse input from switches to alphabet and shows on the screen and vice versa. This project was selected as 4th best project in class of 113 students in digital design course. Written with VHDL.
![Screenshot of web app](/assets/img/algotrading.webp)
Algorithmic trader implemented with Python
- Stock trading algorithm with Python using basic indicators such as moving averages, some of those were partially successful. Also built stock scoring script based on recent news from internet.
- Utilized Yahoo APIs and variety of indicators from different sources such as Moving Averages etc.
Skills
Languages and Databases
![](/assets/img/python-logo-1-300x300.jpg)
![](/assets/img/p-80-java.webp)
![](/assets/img/mysql-logo-1-300x300.jpg)
![](/assets/img/postgresql-logo.png)
![](/assets/img/1200px-C_Programming_Language.svg.png)
![](/assets/img/shell-logo-1-300x300.jpg)
Libraries
![](/assets/img/numpy-logo-1-500x500.jpg)
![](/assets/img/pandas-logo-2-500x500.jpg)
![](/assets/img/sk-learn-logo-1-500x500.jpg)
![](/assets/img/matplotlib-logo-1-500x500.jpg)
![](/assets/img/xgboost.png)
![](/assets/img/opencv-logo-1-500x500.jpg)
![](/assets/img/Apache_POI_project_logo_(2018).svg.png)
Frameworks
![](/assets/img/django-logo.webp)
![](/assets/img/OG-Spring.png)
![](/assets/img/giphy.gif)
Other
![](/assets/img/git.png)
![](/assets/img/aws.png)
![](/assets/img/Docker_Temporary_Image_Google_Blue_1080x1080_v1.webp)