Gaurab Regmi

Key Skills

Software

Mindrift

Toloka

Top Subject Matter

Physics

Data Annotation

STEM

Top Data Types

Audio

Video

Document

Top Task Types

Transcription

Polygon

Bounding Box

Cuboid

Text Generation

Question Answering

RLHF

Freelancer Overview

As a Physics professional and AI Training Specialist, I bring a unique combination of high-level scientific rigor and technical data expertise to OpenTrain AI. My experience at Mindrift and Toloka has focused on enhancing LLM performance through precise data annotation, audio segmentation, and the evaluation of complex NLP datasets. I am particularly skilled at performing ground-truth verification for technical content, ensuring that model outputs are not only linguistically fluent but also scientifically accurate. What sets me apart is my ability to bridge the gap between theoretical physics and machine learning. With an M.Sc. in Physics and proficiency in Python (NumPy, SciPy), I can provide expert-level RLHF (Reinforcement Learning from Human Feedback) for advanced STEM queries that generalist annotators cannot handle. Whether I am evaluating agent-based workflows or labeling intricate scientific datasets, I apply the same analytical precision required in physics to ensure the highest quality training data for next-generation models.

Labeling Experience

AI Training Specialist (Freelance)

MindriftAudioTranscription

As an AI Training Specialist (Freelance) at Mindrift, I performed data annotation, transcription, and evaluation tasks for machine learning models. My work contributed to improving the dataset quality for NLP and speech recognition systems. I regularly worked on high-volume microtasks with precision, following strict quality guidelines and model improvement protocols. • Audio transcription and segmentation for speech recognition datasets • Labeled and annotated spoken content for NLP model training • Ensured task consistency and accuracy through detailed evaluation • Collaborated with team standards to boost model performance

As an AI Training Specialist (Freelance) at Mindrift, I performed data annotation, transcription, and evaluation tasks for machine learning models. My work contributed to improving the dataset quality for NLP and speech recognition systems. I regularly worked on high-volume microtasks with precision, following strict quality guidelines and model improvement protocols. • Audio transcription and segmentation for speech recognition datasets • Labeled and annotated spoken content for NLP model training • Ensured task consistency and accuracy through detailed evaluation • Collaborated with team standards to boost model performance

2026 - Present

Education

Work History