At Turing, I engineer and validate benchmark tasks for frontier AI models. I debug task logic, execution environments, test suites, and automated verifiers to support reliable, reproducible evaluation.
I also develop automated workflows to assess model-generated outputs, analyze failure patterns across runs, and calibrate task difficulty using cross-source reasoning and edge-case analysis.
At JungleWorks, I contributed to production web applications by building responsive, reusable UI components with React.js and JavaScript. I also worked on debugging, performance optimization, and code integration in an Agile environment.
My projects include a full-stack AI website generator, a grocery search and checkout system, and a Vision Transformer-based pneumonia detection system. These projects involved building page-generation pipelines, transactional order flows, and image-preprocessing and model-training pipelines.
