Applied AI · OCR · cloud
Attendance from difficult images
Built an OCR attendance application with Azure AI Vision and GPT-4o for low-contrast employee photographs. It reached 98% accuracy and supported about 100 users.
PythonAzureGPT-4o
Selected work · software · research
I like turning hard problems
into working software.
I’m interested in applied AI, backend engineering, data, and the computational tools that help people make better decisions.
Scientific computing · CMS/CERN
Finding structure in collisions
Modified a large C++ analysis program and wrote Python, ROOT, and Bash tooling for heavy-ion collision data, systematic uncertainties, and multiparticle flow studies.
C++PythonROOT
NLP · data engineering
Keeping useful context in the system
Built a multilingual, multi-client NER service for sanitizing sensitive information, and contributed metadata generation to improve retrieval context for LLM prompts.
PythonAWSNLP
Background
Physics taught me to look for the model underneath the noise.
Before software engineering, I studied Engineering Physics at IIT Roorkee and Physics at IIT Madras. I’m now pursuing an MS in Software Engineering at San José State University.