Alex Mugo

Key Skills

Software

CVAT

Labelbox

Label Studio

Top Subject Matter

LLM evaluation

Rlhf Domain Expertise

Prompt Engineering

Top Data Types

Text

Image

Video

Top Task Types

RLHF

Bounding Box

Computer Programming Coding

Segmentation

Freelancer Overview

Remote AI Trainer / LLM Evaluator. Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and CVAT. Education includes Master of Science, University College London and Bachelor of Science, University of Manchester. AI-training focus includes data types such as Text and Image and labeling workflows including RLHF, Evaluation, and Rating.

Labeling Experience

Remote AI Trainer / LLM Evaluator

TextRLHF

As a Remote AI Trainer and LLM Evaluator, I evaluated over 15,000 large language model (LLM) responses, focusing on reasoning, coding, and factual verification tasks. I delivered RLHF feedback contributing to notable reductions in model hallucination rates. I regularly designed adversarial and edge-case prompts for stress-testing model capabilities. • Conducted output ranking and benchmarking to enhance evaluation consistency • Identified recurring reasoning failures to improve model alignment • Increased quality of training signals through systematic evaluation • Emphasized model behavior analysis and reliable output generation

As a Remote AI Trainer and LLM Evaluator, I evaluated over 15,000 large language model (LLM) responses, focusing on reasoning, coding, and factual verification tasks. I delivered RLHF feedback contributing to notable reductions in model hallucination rates. I regularly designed adversarial and edge-case prompts for stress-testing model capabilities. • Conducted output ranking and benchmarking to enhance evaluation consistency • Identified recurring reasoning failures to improve model alignment • Increased quality of training signals through systematic evaluation • Emphasized model behavior analysis and reliable output generation

2023 - Present

Data Annotation Specialist

CVATImageBounding Box

I annotated over 10,000 images, videos, and audio samples using CVAT, Labelbox, and Label Studio for computer vision tasks. My work included applying bounding boxes, segmentation, and classification labels for complex visual AI datasets. I maintained rigorous annotation quality standards through systematic QA workflows. • Enforced labeling guidelines and validation checks • Supported preparation of datasets for multimodal AI systems • Applied a variety of annotation types across images and video • Worked with industry-standard annotation tools and platforms

I annotated over 10,000 images, videos, and audio samples using CVAT, Labelbox, and Label Studio for computer vision tasks. My work included applying bounding boxes, segmentation, and classification labels for complex visual AI datasets. I maintained rigorous annotation quality standards through systematic QA workflows. • Enforced labeling guidelines and validation checks • Supported preparation of datasets for multimodal AI systems • Applied a variety of annotation types across images and video • Worked with industry-standard annotation tools and platforms

2022 - 2023

Machine Learning Data Analyst

Text

As a Machine Learning Data Analyst, I built and processed over one million data points for natural language processing training pipelines. I improved dataset quality through established cleaning, normalization, and validation workflows. I developed benchmarks to measure model accuracy and reasoning. • Produced reports highlighting dataset gaps and model weaknesses • Supported analytical evaluation of NLP model performance • Ensured data preparation for robust machine learning training • Maintained focus on dataset accuracy and reliability

As a Machine Learning Data Analyst, I built and processed over one million data points for natural language processing training pipelines. I improved dataset quality through established cleaning, normalization, and validation workflows. I developed benchmarks to measure model accuracy and reasoning. • Produced reports highlighting dataset gaps and model weaknesses • Supported analytical evaluation of NLP model performance • Ensured data preparation for robust machine learning training • Maintained focus on dataset accuracy and reliability

2021 - 2023

Education

Work History