Alade Tunde Adewunmi

Key Skills

Software

LabelImg

Scale AI

Supervisely

Surge AI

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Classification

Segmentation

Translation Localization

Freelancer Overview

I am a detail-oriented data labeling and AI training specialist with hands-on experience annotating diverse datasets across image, text, audio, and video formats. I’ve contributed to high-impact projects involving natural language processing (NLP), object detection, sentiment analysis, and LLM evaluation—especially in multilingual contexts like Spanish. My work includes bounding box and polygon segmentation, NER tagging, sentiment classification, and model evaluation, with a strong focus on consistency, quality control, and domain-specific accuracy. I’ve used tools across a range of industries including NLP, retail, and education. I’m particularly skilled in following detailed annotation guidelines, collaborating with QA teams to improve labeling standards, and adapting to new ontologies and tools quickly. What sets me apart is my ability to handle nuanced tasks—like evaluating language models in Spanish—with a strong sense of linguistic and cultural awareness.

Labeling Experience

LLM Output Evaluation – Spanish

Surge AITextClassificationQuestion Answering

This project involved evaluating the outputs of a multilingual large language model (LLM), specifically in Spanish. Tasks included prompt + response rating, verifying factual correctness, fluency, and relevance of generated text, and performing classification on model behavior across question-answering tasks. I was also involved in suggesting prompt improvements and writing alternative completions. The dataset included over 10,000 samples across varied domains including customer service, general knowledge, and conversational AI. Strict quality control measures were applied, including inter-annotator agreement checks and gold standard reviews.

This project involved evaluating the outputs of a multilingual large language model (LLM), specifically in Spanish. Tasks included prompt + response rating, verifying factual correctness, fluency, and relevance of generated text, and performing classification on model behavior across question-answering tasks. I was also involved in suggesting prompt improvements and writing alternative completions. The dataset included over 10,000 samples across varied domains including customer service, general knowledge, and conversational AI. Strict quality control measures were applied, including inter-annotator agreement checks and gold standard reviews.

2023 - 2024

Audio Transcription and Speaker Tagging for Call Center Data

LabelboxAudioClassificationEmotion Recognition

Transcribed and annotated large volumes of customer support audio recordings for use in training automatic speech recognition (ASR) systems. Tasks included converting spoken dialogue into accurate text, tagging speaker turns for diarization, identifying emotional tone (e.g., frustration, urgency), and classifying call outcomes. The project involved over 5,000 audio files in English and Pidgin English, often under noisy conditions. Adhered to strict timestamp alignment rules and quality checks, with accuracy benchmarks above 95% and peer reviews for consistency.

Transcribed and annotated large volumes of customer support audio recordings for use in training automatic speech recognition (ASR) systems. Tasks included converting spoken dialogue into accurate text, tagging speaker turns for diarization, identifying emotional tone (e.g., frustration, urgency), and classifying call outcomes. The project involved over 5,000 audio files in English and Pidgin English, often under noisy conditions. Adhered to strict timestamp alignment rules and quality checks, with accuracy benchmarks above 95% and peer reviews for consistency.

2022 - 2023

Education

Work History