Timothy Kozak

Key Skills

Software

Clickworker

CloudFactory

CrowdSource

Google Cloud Vertex AI

Mindrift

OneForma

Surge AI

AWS SageMaker

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Video

Top Task Types

Classification

Computer Programming Coding

Data Collection

Emotion Recognition

Object Detection

Freelancer Overview

I have hands-on experience in data labeling and AI training through a combination of academic projects, professional roles, and AI-focused competitions. My work has involved preparing, cleaning, and annotating structured datasets for machine learning models—particularly in the financial and healthcare sectors. At DTE Analytics, I developed and maintained SQL-driven dashboards and worked closely with cross-functional teams to support strategic decisions through data analysis. I’ve also participated in Datathons where I helped build AI models for illness detection and led data annotation tasks that directly impacted model performance. In addition to industry experience, I hold professional certifications in Machine Learning, AI Engineering, and Power BI, which have equipped me with strong technical skills in data preprocessing, labeling, model evaluation, and visualization. I’m particularly skilled at handling tabular and text data, and I’ve contributed to projects involving LLM evaluation, sentiment analysis, and NLP-based classification tasks. With a strong foundation in statistics, SQL, and Python, and a collaborative mindset, I bring a well-rounded and adaptable approach to AI training data projects.

Labeling Experience

Tabular Data Annotation for Financial Forecasting Models

Aws SagemakerTextEntity Ner ClassificationSegmentation

This was an academic capstone project focused on forecasting stock price movements and business performance using macroeconomic and financial indicators. I was responsible for labeling data points based on market movement categories (e.g., rise, fall, stable) and annotating events based on earnings releases, economic reports, and company news. Labels were used to train a logistic regression and ensemble learning model. The project involved verifying data accuracy, reducing noise in the labels, and ensuring consistent formatting for ML training. I also helped create logic for mapping features to specific model outputs, simulating function calling for LLM-based forecasting assistance.

This was an academic capstone project focused on forecasting stock price movements and business performance using macroeconomic and financial indicators. I was responsible for labeling data points based on market movement categories (e.g., rise, fall, stable) and annotating events based on earnings releases, economic reports, and company news. Labels were used to train a logistic regression and ensemble learning model. The project involved verifying data accuracy, reducing noise in the labels, and ensuring consistent formatting for ML training. I also helped create logic for mapping features to specific model outputs, simulating function calling for LLM-based forecasting assistance.

2024 - 2024

Illness Severity Classification and Medical Data Labeling – DTE Consultancy Datathon

Aws SagemakerTextClassificationObject Detection

This project was part of the DTE Consultancy Datathon in 2023, focused on developing a predictive model to assess illness severity using structured healthcare data. I was responsible for labeling medical and patient records to prepare the dataset for supervised learning. Tasks included categorizing patient cases based on symptoms, outcomes, and severity levels, as well as tagging clinical terms using entity recognition techniques. I helped design the labeling schema, ensured consistency in class assignments, and validated labeled data to maintain high annotation quality. Additionally, I contributed to the evaluation and refinement of labeled data through iterative model feedback loops. Our final dataset supported the development of a model that accurately predicted illness severity, and the project was recognized by judges for its practical healthcare application.

This project was part of the DTE Consultancy Datathon in 2023, focused on developing a predictive model to assess illness severity using structured healthcare data. I was responsible for labeling medical and patient records to prepare the dataset for supervised learning. Tasks included categorizing patient cases based on symptoms, outcomes, and severity levels, as well as tagging clinical terms using entity recognition techniques. I helped design the labeling schema, ensured consistency in class assignments, and validated labeled data to maintain high annotation quality. Additionally, I contributed to the evaluation and refinement of labeled data through iterative model feedback loops. Our final dataset supported the development of a model that accurately predicted illness severity, and the project was recognized by judges for its practical healthcare application.

2023 - 2023

Education

Work History