Sanchit B.

Key Skills

Software

Other

Labelbox

Top Subject Matter

Conversational AI

ASR transcription

LLM evaluation (English/Hindi)

Top Data Types

Image

Video

Text

Audio

Document

Top Task Types

Bounding Box

Tracking

Text Summarization

RLHF

Audio Recording

Transcription

Segmentation

Prompt + Response Writing (SFT)

Data Collection

Freelancer Overview

AI / LLM Annotator & Evaluator (Freelance) | Innodata. Brings 4+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Innodata Internal Tools, Other, and Labelbox. Education includes Master of Science, WorldQuant University (2025) and Bachelor of Arts, Ambedkar University, Delhi (2023). AI-training focus includes data types such as Audio, Document, and Text and labeling workflows including Audio Recording, Evaluation, and Rating.

Labeling Experience

AI/LLM Analyst - Innodata

LabelboxTextSegmentationRLHF

You evaluate and annotate AI and LLM-generated outputs, assessing factual accuracy, response quality, and alignment with user intent. You design and develop prompts to stress-test AI systems and identify edge cases across reasoning, perception, and language understanding. You analyze and compare model responses using structured rubrics to provide feedback that refines model behavior. • Conduct quality evaluation of AI/LLM outputs • Build prompt suites for robustness testing • Use rubrics and scoring to compare responses • Collaborate to maintain consistent quality for generative AI pipelines

2026 - Present

AI / LLM Annotator & Evaluator (Freelance) | Innodata

AudioAudio RecordingSegmentation

Executed Speech-to-Speech (S2S) audio evaluation for conversational AI, focusing on naturalness, audio quality, and persona likeness. Completed ASR transcription with timestamp alignment and speaker diarisation across diverse accents and domains. Assessed multimodal audio-related tasks as part of scalable, zero-error annotation workflows for LLM and speech system quality. • S2S audio evaluation (naturalness, quality, persona likeness) • ASR transcription with timestamps and speaker diarisation • Stress-tested language understanding across English and Hindi prompts • Supported internal certification requirements for annotation fundamentals

2026 - Present

Subject Matter Expert – Research & Analytics | Vaidik Eduservices Pvt. Ltd.

OtherDocumentData Collection

Researched and structured economics modules to create high-quality content usable for consistent dataset creation and annotation quality. Applied rigorous QC practices to ensure zero-error research outputs under tight deadlines. Supported analytics work that improved downstream interpretation quality for AI/data pipelines. • Structured 50+ economics modules for dataset consistency • Zero-error outputs via rigorous quality control • Research/analytics contributions supporting AI dataset reliability • Deadline-driven execution supporting data work

2023 - 2025

I worked on a live Speech-to-Speech (S2S) AI model, where I trained and evaluated speech models across different scenari

I worked on a live Speech-to-Speech (S2S) AI model, where I trained and evaluated speech models across different scenarios. My role involved comparing model responses and rating them across multiple categories such as accuracy, relevance, naturalness, and overall performance to determine which model performed better.

Not specified

Education

Work History