At CodeContent, I develop computer vision pipelines using Roboflow, from annotation and augmentation through training and inference. I also publish reproducible implementations on the Roboflow Blog.
I benchmark AI models and variants with Weights & Biases and Weave, comparing performance metrics and execution traces to identify strengths, limitations, and failure patterns.
At DigitalX, I designed and built a production Survey Management System backend that supported 2,000+ users. I handled its architecture, secure REST APIs, database schema, migrations, and Docker containerization.
In my projects, I built ScolioCare, a DenseNet-121 system for scoliosis detection that achieved 92.6% clinical test accuracy, and Faultline, a platform for evaluating LLM agents against production-incident simulations.

