Kumar Shalibhadra Mehta

I’m a B.Tech in Computer Science Engineering student from Vellore Institute of Technology and I work best where software engineering meets AI. I’ve contributed full-stack features and APIs that turn complex risk signals into actionable, real-time views for end users.

In my SDE Intern role at State Bank Global IT Centre (Software Factory), I helped build PRISM, an enterprise early-warning system for clients seeking 30+ crore loans. I engineered a data ingestion and compression pipeline (2,300+ raw data points into 236 features), integrated ML/DL models for high-risk account flagging, and delivered monitoring dashboards with end-to-end REST APIs using AngularJS and Spring Boot.

My projects reflect my focus on practical AI pipelines: a GenAI Pipeline for hybrid Rule–LLM compliance using LLaMA 3.3-70B via Groq, and a Zero-Day Attack Detection system using a hybrid IDS (autoencoder + Random Forest) with CICIDS2017. I also developed an Agentic RAG system for semantic retrieval, fact verification, and adaptive response generation using LangChain and FastAPI, improving accuracy while reducing hallucinations.

I enjoy building user-facing analytics too—like a digital cashbook/ledger in R with Shiny and Plotly—and I like designing experiences that feel fast and reliable, from interactive maps (React) to audit-friendly systems (SQLite + Streamlit). I’m excited to keep shipping production-minded AI and software that makes decisions clearer, faster, and safer.