AI / ML
Model evaluation, failure discovery and evidence-backed prediction.
What that involves
Model evaluation
Investigating classifier behavior beyond aggregate accuracy.
Failure discovery
Searching for data subgroups with statistically supported weaknesses.
Statistical validation
Hypothesis tests, FDR correction and holdout checks.
Regression analysis
Comparing model versions using preserved failure evidence.
(In practice)
Engineering grounded in working projects.
Explore the implementation, technical decisions and known limitations in my project case studies.
Questions about AI / ML
The Work section links to project case studies and confirmed public GitHub repositories. Each case study distinguishes verified implementation from documented plans.
My toolkit includes Python, Java, TypeScript, React, Next.js, FastAPI, Spring Boot, Scikit-learn, MySQL and PostgreSQL. I also use Tree-sitter and NetworkX for repository analysis.
I am studying B.Tech in Computer Science and Engineering (Core) at SRM Institute of Science and Technology, Kattankulathur, with expected graduation in May 2028.
Email ashmitrai12@gmail.com or use the contact page to prepare an email draft. I am based in Chennai, Tamil Nadu, India.