Osaid Ashraf

Key Skills

Software

Appen

Clickworker

Data Annotation Tech

Labelbox

OneForma

Scale AI

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Image

Top Task Types

Action Recognition

Audio Recording

Computer Programming Coding

Data Collection

Freelancer Overview

I'm an AI training specialist skilled in both coding and audio data annotation I've built and optimized Python based pipelines for transcribing diarizing and noise labeling multi speaker recordings boosting throughput by 30% my combined software and audio expertise ensures high quality ready to train data sets for speech models

Labeling Experience

audio emotion and speaker lebeling

Data Annotation TechAudioClassificationTranslation Localization

annotated podcast and conversational audio by segmenting speaker turns and tagging emotions such as happy neutral surprised used label studios waveform interface and scale AIS dashboard to label approximately 150 clips approximately 3 minutes each with multi pass review ensuring greater than 90% into annotator agreement this workflow improved data set reliability for emotion detection and diarization models

annotated podcast and conversational audio by segmenting speaker turns and tagging emotions such as happy neutral surprised used label studios waveform interface and scale AIS dashboard to label approximately 150 clips approximately 3 minutes each with multi pass review ensuring greater than 90% into annotator agreement this workflow improved data set reliability for emotion detection and diarization models

2024

coding expert

Scale AIComputer Code ProgrammingComputer Programming Coding

What are you doing evaluated and ranked AI generated Python And JavaScript code crafted coding prompts identified bugs and provided concise feedback completed approximately 200 plus tasks with 95% plus quality accuracy directly enhancing model learning quality

What are you doing evaluated and ranked AI generated Python And JavaScript code crafted coding prompts identified bugs and provided concise feedback completed approximately 200 plus tasks with 95% plus quality accuracy directly enhancing model learning quality

2024

Education

Work History