Yusuf Dhikrullah

Key Skills

Software

AWS SageMaker

CVAT

Data Annotation Tech

Labelbox

Prodigy

Scale AI

SuperAnnotate

Toloka

Telus

V7 Labs

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Audio Recording

Data Collection

Entity Ner Classification

Segmentation

Text Summarization

Freelancer Overview

I am an AI Data Annotation Specialist who has practical experience in training and fine-tuning AI models by text, audio, and video labeling. I have experience with prompt engineering, intent classification, and fact-checking to produce high-quality, contextually appropriate outputs using large language models (LLMs). I have been engaged with projects about data curation, annotation guidelines, and quality control of NLP and generative AI applications. I also know how to use the best annotation software, including Labelbox, SuperAnnotate, and Prodigy, and am very attentive and capable of following complicated directions. Being a researcher, content producer, and technical content writer can better equip me with the ability to provide specific annotations and feedback that can lead to improved performance of AI models. With the willingness to create AI systems, I possess analytical skills, flexibility, and determination to provide correct and well-organized information to future AI programs.

Labeling Experience

Text Annotation for NLP Model Training

LabelboxTextEntity Ner ClassificationClassification

Labeled and annotated more than 5,000 pieces of text in an NLP project devoted to chatbot and virtual assistant performance. The problems were intent classification, named entity recognition (NER) and question-answer labeling. Did quality checks and gave structured feedback to achieve 95 percent or higher accuracy.

Labeled and annotated more than 5,000 pieces of text in an NLP project devoted to chatbot and virtual assistant performance. The problems were intent classification, named entity recognition (NER) and question-answer labeling. Did quality checks and gave structured feedback to achieve 95 percent or higher accuracy.

2023 - 2023

Image Annotation for Computer Vision Model

LabelboxImageBounding BoxPolygon

Labeled 1,500+ product and object pictures in an AI vision application in retail and e-commerce. Bounding boxes and polygon segmentation to detect and categorize items at high accuracy in automated inventory and search solutions. Assured consistency and accuracy of large sets of data on a restricted timeline.

Labeled 1,500+ product and object pictures in an AI vision application in retail and e-commerce. Bounding boxes and polygon segmentation to detect and categorize items at high accuracy in automated inventory and search solutions. Assured consistency and accuracy of large sets of data on a restricted timeline.

2024 - 2025

LLM Evaluation & Reinforcement Learning with Human Feedback

Scale AITextQuestion AnsweringText Generation

Tested 2,000+ responses generated by LLM on accuracy, tone and adherence to user intent. Fact-checking, red-teaming bias detection and formal feedback to optimize the model. Install immediate optimization on Supervised Fine-Tuning (SFT), and introduce context and relevance of input to conversational AI software.

Tested 2,000+ responses generated by LLM on accuracy, tone and adherence to user intent. Fact-checking, red-teaming bias detection and formal feedback to optimize the model. Install immediate optimization on Supervised Fine-Tuning (SFT), and introduce context and relevance of input to conversational AI software.

2024 - 2024

Video Annotation for Object Detection in AI Vision Models

CVATVideoBounding BoxObject Detection

Labeled 200+ video segments as part of an AI vision model using objects and frame by frame tracking. Bounding boxes and action recognition tags that are used in detecting vehicles, pedestrians, and objects in dynamic situations. Strict quality guidelines to achieve 99 percent accuracy in labeled information to train autonomous navigation systems.

Labeled 200+ video segments as part of an AI vision model using objects and frame by frame tracking. Bounding boxes and action recognition tags that are used in detecting vehicles, pedestrians, and objects in dynamic situations. Strict quality guidelines to achieve 99 percent accuracy in labeled information to train autonomous navigation systems.

2024 - 2024

Audio Annotation for Speech Recognition Model

SuperannotateAudioEmotion RecognitionEvaluation Rating

Transcribed audio, labelling speakers, and tagging the emotions of 300 or more audio files to create a speech-to-text system. Assured uniform marking of files, noise filtering principles and ensured high precision of the ASR (Automatic Speech Recognition) model training. Provided high quality data to enhance the functionality of voice assistants.

Transcribed audio, labelling speakers, and tagging the emotions of 300 or more audio files to create a speech-to-text system. Assured uniform marking of files, noise filtering principles and ensured high precision of the ASR (Automatic Speech Recognition) model training. Provided high quality data to enhance the functionality of voice assistants.

2023 - 2023

Education

Work History