Things I've built — from multi-agent AI platforms to 5G network systems and computer vision models.
Production-grade multi-agent platform. A LangGraph StateGraph orchestrates dual Llama3 scorer agents evaluating GitHub profiles across 8 rubrics with consensus scoring, pgvector semantic search, and smart cache invalidation.
Anonymous reporting platform verifying identity while preserving anonymity. Uses Google Gemini AI to neutralise toxic language and route complaints intelligently.
Multimodal AI pipeline combining MobileNetV2 (CNN) for facial analysis and LSTM networks processing MFCC audio features. Deployed via Flask.
Web app automating User Equipment and gNodeB operations with Open5GS core integration and UERANSIM simulation for full 5G network lifecycle management.
Productivity dashboard built during Edunet internship. Features task scheduling, progress analytics, Pomodoro timer and RESTful APIs for real-time updates.
ML model using sensor anomaly analysis to detect early water leakage. Trained on IoT sensor datasets using Scikit-learn pipelines with high precision.