Isaac Mugenya

Key Skills

Software

Appen

Other

Toloka

Top Subject Matter

Swahili Speech Recognition

LLM Training (Swahili and English)

AI Model Training (General)

Top Data Types

Audio

Text

Image

Top Task Types

Transcription

Classification

Freelancer Overview

Swahili Language Specialist. Core strengths include Appen, Other, and Toloka. Education includes Bachelor of Science, School of Applied Sciences, Meru University (2024). AI-training focus includes data types such as Audio, Text, and Image and labeling workflows including Transcription, Evaluation, and Rating.

Labeling Experience

Swahili Language Specialist

AppenAudioTranscription

As a Swahili Language Specialist at CrowdGen (Appen), I transcribed Swahili audio files for training AI speech recognition models. My work involved evaluating machine-generated transcriptions for grammatical accuracy, fluency, and cultural appropriateness. I strictly followed project-specific guidelines and contributed feedback to improve annotation processes. • Labeled speakers, timestamped audio segments, and annotated non-speech sounds per instructions • Verified adherence to Swahili writing systems and standards • Maintained consistent daily output and met quality benchmarks • Flagged ambiguous instructions and collaborated with project leads

As a Swahili Language Specialist at CrowdGen (Appen), I transcribed Swahili audio files for training AI speech recognition models. My work involved evaluating machine-generated transcriptions for grammatical accuracy, fluency, and cultural appropriateness. I strictly followed project-specific guidelines and contributed feedback to improve annotation processes. • Labeled speakers, timestamped audio segments, and annotated non-speech sounds per instructions • Verified adherence to Swahili writing systems and standards • Maintained consistent daily output and met quality benchmarks • Flagged ambiguous instructions and collaborated with project leads

2024 - Present

Data Annotator

TolokaImageClassification

As a Data Annotator with Toloka AI, I labeled images, text, and audio for AI model development across various tasks. My responsibilities included following detailed annotation guidelines and adapting to new task requirements. I ensured high accuracy and independently self-reviewed all work prior to submission. • Labeled multiple data formats: images, text, and audio • Consistently met quality and volume benchmarks • Communicated clarifying questions to supervisors when needed • Adapted quickly to diverse annotation formats

As a Data Annotator with Toloka AI, I labeled images, text, and audio for AI model development across various tasks. My responsibilities included following detailed annotation guidelines and adapting to new task requirements. I ensured high accuracy and independently self-reviewed all work prior to submission. • Labeled multiple data formats: images, text, and audio • Consistently met quality and volume benchmarks • Communicated clarifying questions to supervisors when needed • Adapted quickly to diverse annotation formats

2023 - Present

AI Data Annotator

OtherText

At Outlier AI, I annotated Swahili and English text and conversational data used to fine-tune large language models. I evaluated AI-generated responses for fluency, correctness, and cultural fit in Swahili contexts. I rigorously QA-checked peer annotation and supported guideline improvements through identifying edge cases. • Created and refined question-answer pairs for real-world Swahili use cases • Performed quality assurance checks on datasets • Provided feedback to ensure annotation consistency • Delivered work on time and across diverse task types

At Outlier AI, I annotated Swahili and English text and conversational data used to fine-tune large language models. I evaluated AI-generated responses for fluency, correctness, and cultural fit in Swahili contexts. I rigorously QA-checked peer annotation and supported guideline improvements through identifying edge cases. • Created and refined question-answer pairs for real-world Swahili use cases • Performed quality assurance checks on datasets • Provided feedback to ensure annotation consistency • Delivered work on time and across diverse task types

2023 - Present

Education

Work History