Kenan Kerem Öktener

Key Skills

Software

AWS SageMaker

CVAT

Data Annotation Tech

Label Studio

Prodigy

Scale AI

Top Subject Matter

"Autonomous Systems"

"Software Development & AI Integration"

"Natural Language Processing (NLP)"

Top Data Types

Computer Code Programming

Image

Text

Top Task Types

Classification

Computer Programming/Coding

Data Collection

Fine-tuning

RLHF

Freelancer Overview

I have extensive experience in data labeling and AI training, specializing in creating high-quality datasets for natural language processing (NLP), computer vision, and audio-video models. My expertise includes annotating diverse data types, designing culturally localized prompts, and curating multilingual datasets to improve model performance and relevance. I am proficient in using advanced tools like Label Studio, Prodigy, and AWS SageMaker Ground Truth, as well as integrating labeled data into machine learning workflows. In addition to labeling and annotation, I have significant experience evaluating AI models based on Localization, Instruction Following, Truthfulness, and Writing Quality, providing actionable feedback to optimize performance. My skills in Python and Java allow me to streamline workflows and automate data preparation processes, ensuring efficiency and scalability. This blend of technical expertise and hands-on experience sets me apart, enabling me to deliver robust, user-centric solutions for AI-driven projects.

Labeling Experience

Code Annotation for AI-Driven Code Completion Models

Scale AIComputer Code ProgrammingComputer Programming Coding

Annotated a dataset of code snippets to train AI models for code completion and error detection. Tasks included labeling function definitions, parameter types, and logical structures across multiple programming languages like Python, Java, and JavaScript. Designed and implemented a script to automate the extraction and tagging of key code components, ensuring consistent formatting and structure. Collaborated with a cross-functional team to refine annotation guidelines and integrate labeled data into the client’s machine learning pipeline. The project enhanced the AI model’s ability to predict accurate code completions and identify potential bugs, reducing developer effort by 25%.

Annotated a dataset of code snippets to train AI models for code completion and error detection. Tasks included labeling function definitions, parameter types, and logical structures across multiple programming languages like Python, Java, and JavaScript. Designed and implemented a script to automate the extraction and tagging of key code components, ensuring consistent formatting and structure. Collaborated with a cross-functional team to refine annotation guidelines and integrate labeled data into the client’s machine learning pipeline. The project enhanced the AI model’s ability to predict accurate code completions and identify potential bugs, reducing developer effort by 25%.

2024 - 2024

Image Annotation for Object Detection in Autonomous Vehicles

Label StudioImageBounding Box

Annotated a dataset of high-resolution traffic images to train object detection models for autonomous vehicles. Tasks included drawing precise bounding boxes around objects such as pedestrians, vehicles, and traffic signs while adhering to detailed labeling guidelines. Ensured consistency and accuracy through inter-annotator agreement checks and quality assurance processes. The project improved object detection model performance by enhancing the accuracy of real-world object recognition in various lighting and weather conditions.

Annotated a dataset of high-resolution traffic images to train object detection models for autonomous vehicles. Tasks included drawing precise bounding boxes around objects such as pedestrians, vehicles, and traffic signs while adhering to detailed labeling guidelines. Ensured consistency and accuracy through inter-annotator agreement checks and quality assurance processes. The project improved object detection model performance by enhancing the accuracy of real-world object recognition in various lighting and weather conditions.

2024 - 2024

Multilingual Sentiment Analysis Dataset Annotation

Label StudioTextEntity Ner ClassificationClassification

Annotated multilingual text samples for a sentiment classification project using Label Studio. Tasks involved classifying text into categories such as positive, neutral, and negative while maintaining cultural and linguistic accuracy. Quality measures included achieving 95% inter-annotator agreement and adhering to strict labeling guidelines to ensure consistency and reliability. This dataset significantly improved the client’s NLP model accuracy, demonstrating the impact of high-quality annotations.

Annotated multilingual text samples for a sentiment classification project using Label Studio. Tasks involved classifying text into categories such as positive, neutral, and negative while maintaining cultural and linguistic accuracy. Quality measures included achieving 95% inter-annotator agreement and adhering to strict labeling guidelines to ensure consistency and reliability. This dataset significantly improved the client’s NLP model accuracy, demonstrating the impact of high-quality annotations.

2023 - 2024

Education

Work History