Pushkar Workspace

Key Skills

Software

Label Studio

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Classification

Question Answering

Emotion Recognition

Action Recognition

Text Summarization

Freelancer Overview

I am an AI Engineer with hands-on experience working with complex, real-world biomedical datasets—including ClinVar, TCGA, and WHO data—where high-quality data labeling and annotation were essential to building robust machine learning models. My background in healthcare AI and medical imaging has given me a deep understanding of the importance of accurate data preparation, feature engineering, and statistical validation for successful AI outcomes. I have developed and deployed full-stack AI systems for tasks such as brain tumor detection from MRI scans, leveraging deep learning (CNNs, transfer learning) and explainable AI techniques (Grad-CAM, SHAP) to ensure transparency and trustworthiness of results. Skilled in Python, TensorFlow, Keras, and data visualization tools, I am detail-oriented and comfortable translating complex AI outputs into actionable insights for both technical and non-technical stakeholders. My approach prioritizes data quality, reliability, and clear documentation throughout the AI development lifecycle.

Labeling Experience

Text Annotation Learning Project

Label StudioTextClassificationQuestion Answering

Self-initiated practice project to build text annotation skills for AI training. Used free datasets from Hugging Face and Kaggle to: classify text (sentiment, intent, toxicity), perform basic NER, rate/rank AI responses (helpfulness, accuracy, harmlessness), answer questions from text, identify emotions, and write/edit prompts + responses (SFT style). Focused on guideline adherence, consistency, and high accuracy to prepare for professional tasks.

Self-initiated practice project to build text annotation skills for AI training. Used free datasets from Hugging Face and Kaggle to: classify text (sentiment, intent, toxicity), perform basic NER, rate/rank AI responses (helpfulness, accuracy, harmlessness), answer questions from text, identify emotions, and write/edit prompts + responses (SFT style). Focused on guideline adherence, consistency, and high accuracy to prepare for professional tasks.

2025

Education

Work History