Temi Benjamin

Key Skills

Software

CVAT

Data Annotation Tech

Labelbox

Label Studio

OneForma

Scale AI

Appen

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Classification

Prompt Response Writing SFT

Question Answering

Text Generation

Freelancer Overview

I bring extensive expertise in data labeling and annotation, specializing in creating and validating datasets for diverse AI applications. My experience covers text, image, video, and audio annotation, focusing on tasks like intent classification, named entity recognition (NER), sentiment analysis, and action recognition. I have delivered high-quality labeled datasets using tools such as CVAT, Labelbox, and Audacity, ensuring precision and consistency for machine learning models. My skills include advanced tasks like multilingual content moderation, emotion recognition, relationship classification, and contextual adjustments for NLP systems. I excel in refining datasets through rigorous quality assurance, aligning with project goals and linguistic standards. Whether segmenting video for action recognition, annotating audio for speech and sentiment analysis, or reviewing AI-generated content, I prioritize accuracy and usability, consistently enabling high-performance AI solutions.

Labeling Experience

Audio Annotation for Speech Recognition and Sentiment Analysis

AppenAudioSegmentationClassification

Annotated multilingual audio recordings for speech recognition and sentiment analysis tasks. Responsibilities included segmenting audio into speaker turns, transcribing speech accurately, and labeling segments with emotion and intent classifications. Focused on identifying tone, sentiment, and contextual nuances to train advanced speech recognition and sentiment analysis models. Maintained a high standard of quality by cross-verifying annotations and adhering to strict project guidelines. The processed datasets were instrumental in enhancing real-time sentiment detection and transcription accuracy for virtual assistant

Annotated multilingual audio recordings for speech recognition and sentiment analysis tasks. Responsibilities included segmenting audio into speaker turns, transcribing speech accurately, and labeling segments with emotion and intent classifications. Focused on identifying tone, sentiment, and contextual nuances to train advanced speech recognition and sentiment analysis models. Maintained a high standard of quality by cross-verifying annotations and adhering to strict project guidelines. The processed datasets were instrumental in enhancing real-time sentiment detection and transcription accuracy for virtual assistant

2024

LLM Prompt-Response Evaluation & Refinement,

LabelboxTextClassificationQuestion Answering

Reviewed and annotated prompt-response pairs for LLM tasks such as summarization, translation, and Q&A. Evaluated outputs for fluency, relevance, and naturalness to improve model performance.

Reviewed and annotated prompt-response pairs for LLM tasks such as summarization, translation, and Q&A. Evaluated outputs for fluency, relevance, and naturalness to improve model performance.

2024 - 2024

Video Annotation for Action Recognition in Sports Analytics

LabelboxVideoBounding BoxAction Recognition

Annotated video footage of professional sports matches to identify and classify player actions, such as passes, dribbles, tackles, and goals. Applied bounding boxes and tracking to monitor player movements frame by frame and labeled action sequences for machine learning model training. The project involved analyzing complex interactions in real-time scenarios and ensuring precise labeling for high-accuracy action recognition models. Conducted quality checks to verify data consistency and collaborated with the model development team to align annotations with specific use-case requirements.

Annotated video footage of professional sports matches to identify and classify player actions, such as passes, dribbles, tackles, and goals. Applied bounding boxes and tracking to monitor player movements frame by frame and labeled action sequences for machine learning model training. The project involved analyzing complex interactions in real-time scenarios and ensuring precise labeling for high-accuracy action recognition models. Conducted quality checks to verify data consistency and collaborated with the model development team to align annotations with specific use-case requirements.

2023 - 2023

Image Annotation for Object Detection in Autonomous Vehicles

CVATImageBounding BoxEntity Ner Classification

Worked on a large-scale dataset for autonomous vehicle systems, annotating and labeling images with objects such as pedestrians, vehicles, traffic signs, and lane markings. Used bounding boxes and segmentation tools to identify and classify objects while ensuring high spatial accuracy for machine learning model training. Collaborated with the QA team to review annotations for consistency and quality, achieving a 98% approval rate for the dataset. The annotated data was used to improve object detection and decision-making algorithms in real-time driving environments.

Worked on a large-scale dataset for autonomous vehicle systems, annotating and labeling images with objects such as pedestrians, vehicles, traffic signs, and lane markings. Used bounding boxes and segmentation tools to identify and classify objects while ensuring high spatial accuracy for machine learning model training. Collaborated with the QA team to review annotations for consistency and quality, achieving a 98% approval rate for the dataset. The annotated data was used to improve object detection and decision-making algorithms in real-time driving environments.

2022 - 2022

Multilingual Text Annotation for NLP Model Development

AppenTextClassificationQuestion Answering

Contributed to the development of a multilingual NLP system by annotating and reviewing large-scale Hebrew–English text datasets. Tasks included Named Entity Recognition (NER) to identify and classify entities, relationship classification for semantic connections, and contextual text summarization to train the model's comprehension and condensation abilities. Ensured data integrity through rigorous QA processes and adhered to annotation guidelines to maintain linguistic accuracy and contextual consistency. Delivered high-quality annotated datasets within tight deadlines, directly enhancing the performance and reliability of LLM applications.

Contributed to the development of a multilingual NLP system by annotating and reviewing large-scale Hebrew–English text datasets. Tasks included Named Entity Recognition (NER) to identify and classify entities, relationship classification for semantic connections, and contextual text summarization to train the model's comprehension and condensation abilities. Ensured data integrity through rigorous QA processes and adhered to annotation guidelines to maintain linguistic accuracy and contextual consistency. Delivered high-quality annotated datasets within tight deadlines, directly enhancing the performance and reliability of LLM applications.

2021 - 2021

Education

Work History