Geyanth T.

Key Skills

Software

AWS SageMaker

Appen

Clickworker

Google Cloud Vertex AI

Label Studio

Mercor

Roboflow

Scale AI

Toloka

Top Subject Matter

Semantic Code Search

Basketball Object Detection

Top Data Types

Video

Computer Code Programming

Top Task Types

Classification

Object Detection

Freelancer Overview

Semantic Code Search Project - Code Annotation & Model Training. Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include HuggingFace and YOLOv8. Education includes Master of Technology, National Institute of Technology, Tiruchirappalli (2024) and Bachelor of Technology, International Institute of Information Technology, Bhubaneswar (2023). AI-training focus includes data types such as Computer Code, Programming, and Video and labeling workflows including Classification and Object Detection.

Labeling Experience

Basketball Video Object Detection - Annotation & Data Labeling

VideoObject Detection

Designed and executed an automated annotation process for object detection in basketball footage using YOLOv8. Annotated multiple classes including players, ball, and game events in extracted video frames. Performed quality assurance on data labels to enable robust real-time detection models. • Labeled bounding boxes for all tracked objects in each video frame. • Used OpenCV for frame extraction and annotation support. • Validated annotation consistency across the training dataset. • Deployed annotated data into an end-to-end detection pipeline for live inference.

Not specified

Semantic Code Search Project - Code Annotation & Model Training

Classification

Developed a semantic code retrieval system using a bi-encoder model based on CodeBERT, trained on a Kaggle code search dataset. Labeled and classified code snippets for training and evaluating the semantic search model. Evaluated and annotated results to improve cross-language and AST-based model performance. • Labeled code samples with search relevance and topic categories. • Ensured consistent code annotation guidelines throughout the dataset curation. • Contributed to model evaluation using labeled test sets. • Provided feedback on annotation accuracy and process improvements.

Not specified

Education

Work History