Shivang Asthana

Key Skills

Software

Appen

Clickworker

Labelbox

OneForma

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Audio Recording

Prompt Response Writing SFT

Text Generation

Text Summarization

Translation Localization

Freelancer Overview

With a strong analytical background in finance and AI-driven research, I bring specialized expertise in data labeling, AI training data validation, and multilingual prompt engineering. My experience includes: AI Training & Data Validation: Conducted quality checks for map-based AI systems, verifying location accuracy and ensuring precise geospatial data labeling. Trained LLM models on mathematical datasets, enhancing their problem-solving capabilities for quantitative tasks. Multilingual Prompt Engineering: Developed and refined Hindi-language AI prompts for financial literacy platforms (like RupeeKlub), ensuring cultural and linguistic relevance. Translated complex English financial/technical terms into Hindi while maintaining contextual accuracy. Research-Driven Approach: Published AI-focused papers (e.g., "AI-Driven Financial Advisory" in a Scopus-indexed journal) and filed a patent for an AI fraud detection system, demonstrating hands-on experience in training data optimization and model validation. My unique blend of quantitative rigor (GRE Quant: 170/170), multilingual proficiency (fluent English/Hindi), and experience in geospatial/mathematical data labeling positions me to contribute effectively to OpenTrain.ai’s training pipelines.

Labeling Experience

Hindi-English Financial Text Annotation for LLM Fine-Tuning

LabelboxTextText GenerationPrompt Response Writing SFT

Evaluated and ranked 3,000+ AI-generated prompts, real time replies and responses for a multilingual (Hindi-English) chatbot to optimize output quality. rompt Ranking: Assessed and scored AI responses on a 1-5 scale based on: Coherence (logical flow and relevance) Grammar & Syntax (language accuracy in both Hindi and English) Contextual Appropriateness (suitability for financial queries) Cultural Relevance (localized nuances for Indian users)

Evaluated and ranked 3,000+ AI-generated prompts, real time replies and responses for a multilingual (Hindi-English) chatbot to optimize output quality. rompt Ranking: Assessed and scored AI responses on a 1-5 scale based on: Coherence (logical flow and relevance) Grammar & Syntax (language accuracy in both Hindi and English) Contextual Appropriateness (suitability for financial queries) Cultural Relevance (localized nuances for Indian users)

2024

Education

Work History