For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

LoginCreate account
Invite to Job
Lasya Kharvi

Lasya Kharvi

AI Safety Specialist | Prompt Evaluation | Red Teaming

India flagGangolli, India
$15.00/hrExpertAppenMercorMicro1

Key Skills

Software

AppenAppen
MercorMercor
Micro1
OneFormaOneForma
SuperAnnotateSuperAnnotate
TelusTelus
Other
CVATCVAT
Google Cloud Vertex AIGoogle Cloud Vertex AI
TolokaToloka
Internal/Proprietary Tooling

Top Subject Matter

Conversational AI
AI Safety
Data Labeling & Annotation

Top Data Types

TextText
AudioAudio
Geospatial Tiled ImageryGeospatial Tiled Imagery

Top Task Types

Red Teaming
Prompt Response Writing SFT
Mapping
Classification
Audio Recording
Text Summarization
RLHF
Fine Tuning
Transcription
Evaluation Rating
Data Collection
Segmentation
Question Answering

Freelancer Overview

AI training data and model evaluation specialist with 3+ years of hands on experience supporting large scale AI systems through data labeling, prompt creation, response evaluation, safety testing, and quality assurance. I have contributed across high impact projects involving Red Teaming, supervised fine tuning (SFT), prompt and response writing, search relevance, domain quality rating, advertisement classification, translation quality review, multilingual speech and text data collection, and geospatial mapping tasks. My work has focused on improving model accuracy, reasoning, alignment, and trustworthiness by applying detailed guidelines, identifying failure points, and delivering consistent high quality annotations at scale. What sets me apart is a strong combination of analytical thinking, compliance discipline, and real world operations experience. I am skilled in detecting bias, unsafe outputs, jailbreak attempts, prompt injections, misinformation risks, and low quality responses, while providing clear human feedback for model improvement. I work confidently with structured and unstructured datasets including text, audio, geospatial, and tiled imagery. With an additional background in legal operations, financial documentation review, and process management, I bring exceptional attention to detail, fast learning ability, deadline ownership, and reliability in remote project environments. Education includes Bachelor of Commerce from Mangalore University and a Post Graduate Certificate in Computer Applications.

ExpertEnglishHindiKannada

Labeling Experience

Speech & Prompt Data Contributor – Multilingual AI Training

OtherAudioData Collection
I generated and recorded multilingual prompts and speech data in Hindi and English for use in AI model training. My work emphasized linguistic clarity, consistency, and adherence to defined language standards. The resulting dataset supported speech recognition, multilingual understanding, and broader AI development initiatives. • Recorded diverse audio prompts for multi-language AI datasets • Ensured clear pronunciation, intonation, and script compliance throughout • Validated and annotated speech samples for audio training quality • Supported evaluation and curation of audio content for AI system refinement

I generated and recorded multilingual prompts and speech data in Hindi and English for use in AI model training. My work emphasized linguistic clarity, consistency, and adherence to defined language standards. The resulting dataset supported speech recognition, multilingual understanding, and broader AI development initiatives. • Recorded diverse audio prompts for multi-language AI datasets • Ensured clear pronunciation, intonation, and script compliance throughout • Validated and annotated speech samples for audio training quality • Supported evaluation and curation of audio content for AI system refinement

2022 - Present

Translation Quality Reviewer – AI Output Evaluation

OtherTextEvaluation Rating
I reviewed and scored AI-generated translations from Hindi to English for semantic accuracy and fluency. Each translation was evaluated for intent alignment and naturalness using established guidelines. The feedback informed improvements in both translation model outputs and language understanding for AI. • Examined translations for meaning preservation and accurate context • Assessed fluency and naturalness of output sentences • Provided structured ratings for continuous model improvement • Ensured multilingual communication standards were met or exceeded

I reviewed and scored AI-generated translations from Hindi to English for semantic accuracy and fluency. Each translation was evaluated for intent alignment and naturalness using established guidelines. The feedback informed improvements in both translation model outputs and language understanding for AI. • Examined translations for meaning preservation and accurate context • Assessed fluency and naturalness of output sentences • Provided structured ratings for continuous model improvement • Ensured multilingual communication standards were met or exceeded

2022 - Present

Advertisement Evaluation Specialist

OtherTextClassification
I classified and analyzed online advertisements by identifying advertisers, determining intent, and assessing compliance signals. This role enhanced the quality, targeting, and safety of ad content presented to users. My efforts drove increased ad relevance and compliance rates for the platform. • Categorized advertisements based on specified criteria for audience targeting • Identified compliance or policy issues impacting ad approval • Evaluated intent, product claims, and user value for each ad reviewed • Supplied detailed classification data to optimize ad delivery systems

I classified and analyzed online advertisements by identifying advertisers, determining intent, and assessing compliance signals. This role enhanced the quality, targeting, and safety of ad content presented to users. My efforts drove increased ad relevance and compliance rates for the platform. • Categorized advertisements based on specified criteria for audience targeting • Identified compliance or policy issues impacting ad approval • Evaluated intent, product claims, and user value for each ad reviewed • Supplied detailed classification data to optimize ad delivery systems

2022 - Present

Web Content Rater – Domain Quality Evaluation

OtherTextClassification
I evaluated web content in Hindi and English for quality, trustworthiness, and user intent. My work involved identifying spam, misleading pages, and low-value content using detailed evaluation guidelines. These actions helped improve search relevance and information reliability for users. • Scored websites for accuracy, credibility, and content integrity • Applied systematic judgment for domain quality assurance • Identified deceptive or manipulative website behaviors • Provided actionable ratings to calibrate web search algorithms

I evaluated web content in Hindi and English for quality, trustworthiness, and user intent. My work involved identifying spam, misleading pages, and low-value content using detailed evaluation guidelines. These actions helped improve search relevance and information reliability for users. • Scored websites for accuracy, credibility, and content integrity • Applied systematic judgment for domain quality assurance • Identified deceptive or manipulative website behaviors • Provided actionable ratings to calibrate web search algorithms

2022 - Present

Map Data Evaluator - Apple Maps (via CrowdGen)

OtherGeospatial Tiled ImageryTranscription
I conducted guideline-based evaluations of geospatial map data for Apple Maps via CrowdGen. My work focused on reviewing place accuracy, categorization, search relevance, and consistency to ensure map quality. The deliverables directly contributed to mapping reliability and user experience enhancements. • Analyzed place data for spatial accuracy and classification adherence • Evaluated search functionality and location metadata consistency • Flagged and corrected data discrepancies to improve mapping outputs • Followed strict quality and relevance criteria in all evaluations

I conducted guideline-based evaluations of geospatial map data for Apple Maps via CrowdGen. My work focused on reviewing place accuracy, categorization, search relevance, and consistency to ensure map quality. The deliverables directly contributed to mapping reliability and user experience enhancements. • Analyzed place data for spatial accuracy and classification adherence • Evaluated search functionality and location metadata consistency • Flagged and corrected data discrepancies to improve mapping outputs • Followed strict quality and relevance criteria in all evaluations

2022 - Present

Education

M

Mangalore University

Bachelor of Commerce, Commerce

Bachelor of Commerce
2019 - 2019

Work History

M

Mercor

Red Teaming Specialist

Gangolli, Udupi Dist. Kundapura TQ
2026 - Present
C

Crowdgen

Map Data Evaluator - Apple Maps

Gangolli, Udupi Dist. Kundapura TQ
2022 - Present