Lwando Thwala

Key Skills

Software

Appen

Clickworker

CrowdSource

Data Annotation Tech

Other

Internal/Proprietary Tooling

Top Subject Matter

Generative AI

Llms Domain Expertise

AI Ethics

Top Data Types

Document

Image

Text

Audio

Video

Top Task Types

Data Collection

Evaluation/Rating

Fine-tuning

Prompt + Response Writing (SFT)

RLHF

Transcription

Freelancer Overview

AI Model Training & Evaluation Specialist. Core strengths include RWS Group, Outlier, and Rev.com. Education includes Certificate, Cisco Networking Academy (2024). AI-training focus includes data types such as Text and Audio and labeling workflows including RLHF and Transcription.

Labeling Experience

AI Model Training & Evaluation Specialist

TextRLHF

In this position, I performed large-scale AI training, model evaluation, and content moderation for Generative AI and Large Language Models. My responsibilities included evaluating and labeling AI-generated outputs for policy compliance, safety, bias, and ethical risk. I applied RLHF and Supervised Fine-Tuning (SFT) to improve model accuracy and alignment. • Identified and flagged harmful, biased, or non-compliant outputs according to governance guidelines. • Provided structured feedback and detailed annotations to support AI ethics and fairness initiatives. • Ensured compliance and human alignment in model behavior through careful annotation. • Supported the ongoing development of safer, more robust LLMs.

In this position, I performed large-scale AI training, model evaluation, and content moderation for Generative AI and Large Language Models. My responsibilities included evaluating and labeling AI-generated outputs for policy compliance, safety, bias, and ethical risk. I applied RLHF and Supervised Fine-Tuning (SFT) to improve model accuracy and alignment. • Identified and flagged harmful, biased, or non-compliant outputs according to governance guidelines. • Provided structured feedback and detailed annotations to support AI ethics and fairness initiatives. • Ensured compliance and human alignment in model behavior through careful annotation. • Supported the ongoing development of safer, more robust LLMs.

2025 - Present

Project Diamond

Internal Proprietary ToolingImageBounding BoxClassification

Data labelling tasks included object verification, human and object isolation via bounding boxes. Measurement and specification verification and editing.

Data labelling tasks included object verification, human and object isolation via bounding boxes. Measurement and specification verification and editing.

2025

AI Model Training & Evaluation Specialist (Generative AI, NLP, RLHF)

TextRLHF

As an independent contractor, I trained and evaluated Generative AI systems and LLMs utilizing RLHF and Supervised Fine-Tuning workflows. My work focused on prompt–response dataset development, assessment, and improving instruction adherence. I performed output evaluation for accuracy, coherence, safety, and ethical compliance. • Conducted prompt–response writing and review to advance model reasoning capabilities. • Flagged and documented skewed or misleading model behaviors to aid bias mitigation. • Supported NLP-related data labeling such as text classification and summarization. • Maintained high standards for model evaluation rigor and annotation consistency.

As an independent contractor, I trained and evaluated Generative AI systems and LLMs utilizing RLHF and Supervised Fine-Tuning workflows. My work focused on prompt–response dataset development, assessment, and improving instruction adherence. I performed output evaluation for accuracy, coherence, safety, and ethical compliance. • Conducted prompt–response writing and review to advance model reasoning capabilities. • Flagged and documented skewed or misleading model behaviors to aid bias mitigation. • Supported NLP-related data labeling such as text classification and summarization. • Maintained high standards for model evaluation rigor and annotation consistency.

2024 - Present

UHRS

ClickworkerTextQuestion AnsweringText Generation

UHRS offers many AI training tasks which are high in quantity and need a proficient worker to partake in. Audio, Image, Video and mostly Text are included.

UHRS offers many AI training tasks which are high in quantity and need a proficient worker to partake in. Audio, Image, Video and mostly Text are included.

2022

Project WELLS

AppenTextRLHFFine Tuning

The project included multi media such as text, video and audio. Fine-tuning and App store evaluation was included.

The project included multi media such as text, video and audio. Fine-tuning and App store evaluation was included.

2023 - 2024

Show more

Education

Work History