Zenith Cheeli

NetTrace: VASP Attribution Engine (SIH 2026)

Blockchain forensics tool that traces a suspect crypto wallet to the nearest exchange (VASP) across Ethereum, BNB, Polygon and Bitcoin. Builds a transaction graph from live on-chain data, uses bounded BFS to find reachable VASPs and gives explainable confidence scores in a Streamlit investigator dashboard.

GRIDS (Granary Rodent Intrusion Detection System)

Acoustic AI system that detects rodent activity in granaries. MEMS microphones capture sound, MFCC features feed a CNN-LSTM classifier and alerts go out over IoT. Built with Team 42, patent filed and submitted to the MSME 6.0 hackathon.

DDoS Detection in SDN using CNN-LSTM

Binary DDoS classifier on CICIDS2017 that pairs Random Forest feature selection with CNN-LSTM sequence modelling. Reaches 97.47% accuracy against a 99.90% Random Forest baseline and includes a Streamlit dashboard showing predictions and feature contributions.

Stock Price Predictor

LSTM forecasting app that pulls live market data and lets the user pick the ticker, history period and forecast horizon. Scored an MAE of $8.92 and an RMSE of $11.69 on a GOOGL test set.