Frank N.

Key Skills

Software

No software listed

Top Subject Matter

Speech-to-text (audio transcription) dataset labeling for production ML models

Predictive ML dataset creation and labeling for client projects

Operational data validation and monitoring for structured datasets used in technical/ML workflows

Top Data Types

Audio

Text

Image

Top Task Types

Transcription

Classification

Data Collection

Segmentation

Bounding Box

RLHF

Freelancer Overview

Machine Learning Engineer | Violet — Big Idea Africa. Brings 5+ years of professional experience across complex professional workflows, research, and quality-focused execution. Education includes Bachelor of Science, Makerere University (2023). AI-training focus includes data types such as Audio, Text, and Computer Code and labeling workflows including Transcription, Classification, and Data Collection.

Labeling Experience

Machine Learning Engineer - Violet

AudioSegmentation

You prepared and cleaned audio datasets for a speech-to-text model, then labeled and reviewed transcriptions to improve training quality. You handled segmentation and augmentation workflows while applying strict labeling standards and validating output accuracy. Your work required strong understanding of machine learning data preparation, quality assurance practices, and guideline adherence for production-oriented model training. • Prepared and cleaned 200+ hours of audio for speech-to-text training datasets • Managed transcription, segmentation, and data augmentation for model readiness • Applied labeling standards and reviewed outputs to reduce word-error rate • Conducted accuracy checks to maintain quality at scale and meet deadlines

2025 - Present

Machine Learning Engineer | Violet — Big Idea Africa

AudioTranscription

Prepared and labeled 200+ hours of audio for a speech-to-text training dataset, including transcription and segmentation work. Applied strict labeling standards and performed reviews to ensure output accuracy and improve word-error rate. Handled augmentation and preprocessing steps to support high-quality model training and reliable evaluation.• Labeled audio with transcription and segmentation annotations.• Reviewed and validated labeled outputs against guidelines.• Conducted data augmentation and preprocessing for training data.• Tracked improvements to model performance (word-error rate reduced from ~34% to ~15%).

2025 - Present

Software & Blockchain Engineer | Polaris — Big Idea Africa

Data Collection

Aggregated, validated, and monitored large volumes of operational data across 800+ distributed nodes to maintain accurate datasets. Performed validation and consistency checks that support reliable labeled/structured data needed for analytics and ML workflows. Ensured data quality through ongoing monitoring of operational inputs and aggregated outputs.• Aggregated operational data from 800+ distributed nodes.• Validated and monitored data accuracy and consistency.• Supported the creation/maintenance of reliable structured data for modeling/analysis.• Implemented quality controls through routine checks.

2024 - 2025

Data Engineer - Impact Associates And Development Consultants

ImageBounding BoxSegmentation

You collected, preprocessed, and created structured and labeled datasets used to support predictive machine learning across multiple client engagements. You performed data cleaning, validation, and quality checks to ensure datasets were consistent, accurate, and analysis-ready. Your responsibilities required working knowledge of supervised ML data workflows, database/data handling concepts, and careful attention to data quality and edge cases. • Collected and built structured datasets for predictive ML use cases • Preprocessed and labeled data to support downstream modeling • Conducted validation and quality assurance checks across datasets • Supported six client engagements by delivering analysis-ready data

2022 - 2023

Data Engineer | Impact Associates & Development Consultants

TextClassification

Collected, preprocessed, and created structured labeled datasets for predictive machine learning across six client engagements. Performed data cleaning, validation, and quality checks to ensure consistent, analysis-ready labels. Focused on maintaining label integrity and accuracy to support model training and downstream analytics.• Built and structured labeled datasets for predictive ML use cases.• Cleaned and validated data to remove inconsistencies and errors.• Conducted quality assurance checks for label consistency.• Supported six client engagements with accurate labeled outputs.

2022 - 2023

Education

Work History