Pralipta Kumar Sahoo

Key Skills

Software

Appen

Data Annotation Tech

Labelbox

OneForma

Playment

Remotasks

Scale AI

Toloka

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Document

Image

Text

Top Task Types

Prompt Response Writing SFT

Question Answering

Text Generation

Text Summarization

Translation Localization

Freelancer Overview

I have extensive experience in data labeling and AI training data, particularly in the context of mathematical prompts and complex data structures. As a Freelance Math Prompt Engineer at Telus International, I developed, validated, and optimized mathematical problems for AI models, ensuring high-quality, structured datasets that improved model accuracy and performance. My role required strong attention to detail, analytical thinking, and the ability to work with various tools and platforms for data annotation and model training. In addition to this, I have hands-on experience in data transcription and have worked with Google's Data Analytics tools, which have honed my skills in data organization, quality control, and metadata management. These qualifications, coupled with my background in Civil Engineering, make me adept at handling complex data sets and ensuring they are accurately labeled and structured to meet AI training requirements.

Labeling Experience

Audio Annotator

TelusAudioTranslation LocalizationAudio Recording

This project involves audio annotation, where tasks such as transcription, segmentation, classification, and emotion recognition will be carried out on audio data in multiple languages, including Gujarati, French, and English. The primary responsibility includes listening to and annotating audio recordings, providing accurate translations or transcriptions, segmenting the content, and labeling emotional cues as required. The work may also involve recording voice commands for the project. The labeling will be done using software tools provided by Telus, ensuring high-quality output and attention to detail.

This project involves audio annotation, where tasks such as transcription, segmentation, classification, and emotion recognition will be carried out on audio data in multiple languages, including Gujarati, French, and English. The primary responsibility includes listening to and annotating audio recordings, providing accurate translations or transcriptions, segmenting the content, and labeling emotional cues as required. The work may also involve recording voice commands for the project. The labeling will be done using software tools provided by Telus, ensuring high-quality output and attention to detail.

2024

Prompt Response Writer

TelusDocumentPrompt Response Writing SFT

In this project, I served as a Prompt Response Writer specializing in the creation and refinement of AI training data for Large Language Models (LLM) with a focus on Mathematics. My primary responsibility was to generate prompts and their corresponding responses for mathematical concepts, ensuring the AI models received high-quality, contextually accurate data to improve their understanding and problem-solving capabilities. Using Telus as the primary platform, I worked on tasks such as Text Generation, Question Answering, and Function Calling, focusing on developing mathematical questions and answers that would be used to fine-tune the model's capabilities. The goal was to create a comprehensive dataset that improved the model’s performance in mathematical reasoning, from simple arithmetic to complex algebra and calculus problems.

In this project, I served as a Prompt Response Writer specializing in the creation and refinement of AI training data for Large Language Models (LLM) with a focus on Mathematics. My primary responsibility was to generate prompts and their corresponding responses for mathematical concepts, ensuring the AI models received high-quality, contextually accurate data to improve their understanding and problem-solving capabilities. Using Telus as the primary platform, I worked on tasks such as Text Generation, Question Answering, and Function Calling, focusing on developing mathematical questions and answers that would be used to fine-tune the model's capabilities. The goal was to create a comprehensive dataset that improved the model’s performance in mathematical reasoning, from simple arithmetic to complex algebra and calculus problems.

2024

Coder Expert

Scale AIComputer Code ProgrammingComputer Programming Coding

I worked on the Coder Expert project, which focused on data annotation, fine-tuning, and evaluation of Large Language Models (LLMs) for computer programming tasks. My role involved handling various data labeling types, including entity recognition, classification, question answering, text summarization, and fine-tuning. Using Scale AI, I contributed to high-quality annotations and validation processes to enhance LLMs' understanding of coding-related tasks. My work covered areas such as function calling, prompt-response writing (SFT), evaluation/rating, and reinforcement learning from human feedback (RLHF). Additionally, I worked on complex annotation workflows like emotion recognition, tracking, segmentation, and mapping to improve AI-assisted development. Through this project, I gained experience in optimizing data collection and labeling processes to improve coding-related AI systems.

I worked on the Coder Expert project, which focused on data annotation, fine-tuning, and evaluation of Large Language Models (LLMs) for computer programming tasks. My role involved handling various data labeling types, including entity recognition, classification, question answering, text summarization, and fine-tuning. Using Scale AI, I contributed to high-quality annotations and validation processes to enhance LLMs' understanding of coding-related tasks. My work covered areas such as function calling, prompt-response writing (SFT), evaluation/rating, and reinforcement learning from human feedback (RLHF). Additionally, I worked on complex annotation workflows like emotion recognition, tracking, segmentation, and mapping to improve AI-assisted development. Through this project, I gained experience in optimizing data collection and labeling processes to improve coding-related AI systems.

2023 - 2024

Data Analyst

Scale AITextData Collection

In this role, I worked as a Data Analyst for a project focused on enhancing Large Language Models (LLM). My primary responsibility was to analyze and annotate large volumes of text data for training purposes, ensuring the quality and accuracy of the dataset to improve model performance in tasks like Text Generation, Text Summarization, Question Answering, and Classification. Using Scale AI as the primary platform, I worked on diverse data labeling tasks, including Entity Recognition (NER), Sentiment Analysis, and Emotion Recognition, providing high-quality labels to help improve the model’s understanding of text. The project aimed to enhance LLM capabilities by fine-tuning the AI’s ability to process, understand, and generate human-like responses across various domains, including customer service, technical support, and content creation.

In this role, I worked as a Data Analyst for a project focused on enhancing Large Language Models (LLM). My primary responsibility was to analyze and annotate large volumes of text data for training purposes, ensuring the quality and accuracy of the dataset to improve model performance in tasks like Text Generation, Text Summarization, Question Answering, and Classification. Using Scale AI as the primary platform, I worked on diverse data labeling tasks, including Entity Recognition (NER), Sentiment Analysis, and Emotion Recognition, providing high-quality labels to help improve the model’s understanding of text. The project aimed to enhance LLM capabilities by fine-tuning the AI’s ability to process, understand, and generate human-like responses across various domains, including customer service, technical support, and content creation.

2023 - 2024

Image Annotator

TolokaImageBounding BoxPolygon

In this project, I worked as an Image Annotator specializing in AI training data creation for Large Language Models (LLM). The task involved annotating a variety of image datasets to enhance the performance and accuracy of machine learning models across multiple use cases. The types of labeling included Bounding Box, Polygon, Key Points, and Object Detection, among others. I employed Toloka as the main labeling software, ensuring that annotations were precise and aligned with the project's requirements. My contributions focused on ensuring high-quality annotations for LLM training, particularly for tasks like Emotion Recognition, Action Recognition, and Image Classification. The goal was to create a robust training dataset that would improve the AI’s ability to interpret and understand complex visual information.

In this project, I worked as an Image Annotator specializing in AI training data creation for Large Language Models (LLM). The task involved annotating a variety of image datasets to enhance the performance and accuracy of machine learning models across multiple use cases. The types of labeling included Bounding Box, Polygon, Key Points, and Object Detection, among others. I employed Toloka as the main labeling software, ensuring that annotations were precise and aligned with the project's requirements. My contributions focused on ensuring high-quality annotations for LLM training, particularly for tasks like Emotion Recognition, Action Recognition, and Image Classification. The goal was to create a robust training dataset that would improve the AI’s ability to interpret and understand complex visual information.

2020 - 2023

Education

Work History