At Outlier AI (Scale AI), I review contributor work on MCP tool-calling tasks before client delivery, checking accounting and business-database outputs for accuracy and guideline compliance. I progressed from AI Evaluator to Agentic AI Trainer to Project Reviewer.
At Outlier AI, I design prompts, tool-orchestration steps and evaluation criteria for agents working with accounting and business databases. I also author rubrics and classify failure types in multi-step agent trajectories.
At Handshake AI, I create prompts, evaluate model responses and write corrected outputs for RLHF/SFT training data. As a Deep Learning Intern at S.P. College, I built a YOLOv8 traffic-light detection model using a self-collected dataset; I also developed an LSTM stock forecasting project in PyTorch.

