Aniket Swarnokar

Key Skills

Software

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Video

Top Label Types

RLHF

Red Teaming

Prompt Response Writing SFT

Freelancer Overview

Expert AI Evaluation & Alignment Specialist with a Bachelor of Science in Data Science and Analytics. I specialize in high-fidelity Reinforcement Learning from Human Feedback (RLHF) and Adversarial Red Teaming for frontier multimodal models. My core strength lies in bridging the gap between raw data and model safety. I have extensive experience in Temporal & Spatial Technical Auditing, specifically identifying object permanence failures and fluid dynamics violations in generative video. Additionally, I provide high-precision Cross-Lingual Semantic Alignment for English, Hindi, and Bengali datasets to remediate complex linguistic hallucinations. I am proficient in creating schema-valid training data for Function Calling and Agentic Workflows, ensuring models follow structured output requirements with 100% accuracy. I thrive in any labeling stack (Argilla, Scale, custom internal tools) and am available for both high-impact individual tasks and large-scale model evaluation projects.

Labeling Experience

AI Evaluation & Alignment Specialist | Multimodal RLHF & Red Teaming

LabelboxVideoRLHFRed Teaming

Managed the technical alignment and Spatio-Temporal Audit of frontier multimodal models, specifically targeting the remediation of complex failures in generative video. This role involved a rigorous evaluation of Temporal Coherence and Subject Identity Drift, ensuring that character attributes and environmental consistency remained stable across extended frame sequences. I specialized in identifying higher-order "world logic" violations, such as Kinematic Incoherence and failures in Fluid Dynamics, which are often invisible to automated benchmarks but critical for physical grounding.

Managed the technical alignment and Spatio-Temporal Audit of frontier multimodal models, specifically targeting the remediation of complex failures in generative video. This role involved a rigorous evaluation of Temporal Coherence and Subject Identity Drift, ensuring that character attributes and environmental consistency remained stable across extended frame sequences. I specialized in identifying higher-order "world logic" violations, such as Kinematic Incoherence and failures in Fluid Dynamics, which are often invisible to automated benchmarks but critical for physical grounding.

2025

Education

Work History