Key Skills

• 

  Labelbox

  I have hands-on experience using Labelbox to support AI training workflows by annotating and labeling datasets for supervised learning. My role involved tagging edge cases, resolving ambiguous data points, and providing structured feedback to improve model accuracy. Collaborating with engineering and ML teams, I helped enhance labeling guidelines and reduce annotation inconsistencies, enabling scalable and efficient data pipeline operations. I also have experience with data labeling & annotation design.

• 

  Hugging Face

  At Hugging Face, I contributed to building intuitive front-end interfaces that support GenAI workflows for both developers and annotators. I collaborated with cross-functional teams to improve usability of AI tools and integrate content moderation pipelines. Through this work, I became familiar with AI model behavior, helped surface misclassifications, and refined the user experience for reviewing AI-generated outputs. I have experience crafting robust prompt strategies using tools like Hugging Face.

• 

  ChatGPT

  I regularly use ChatGPT to assist in classifying and summarizing unstructured content, validate moderation decisions, and streamline feedback loops. My experience includes prompting GenAI to handle gray-area content, verifying AI outputs, and training models through real-world edge case examples. This enables me to work efficiently in environments where human judgment augments AI intelligence. I built prompt-based UGC moderation flows with fallback mechanisms using ChatGPT.

• 

  Claude

  My work with Claude involves leveraging its contextual reasoning capabilities to moderate nuanced content. I utilize Claude to support annotation decisions, draft policy-aligned responses, and test model responses against ethical and safety benchmarks. This enhances the reliability of moderation processes while minimizing false positives in sensitive domains. I've designed and developed system prompts for AI-based chatbots serving enterprise clients.

• 

  Google Gemini

  Using Gemini, I analyze and classify content across various formats, including images and text, within cross-platform environments. I employ Gemini’s multimodal understanding to assist in edge case detection, data enrichment, and user-generated content moderation. This supports faster iteration cycles and provides high-quality inputs for GenAI system refinement.

• 

  n8n Automation

  I leverage n8n for no-code automation to support AI training and evaluation workflows. My experience includes tagging edge cases and annotating structured data for AI model training, as well as providing feedback to improve the accuracy and fairness of GenAI content classification systems. I also build automated flows with n8n to feed data to labeling teams and record AI performance evaluation results.