Research

AI Research, Technical Writing & Evaluation

First-author papers, research synopses, and technical write-ups spanning LLM systems, multi-agent architectures, computer vision, and applied AI evaluation.

First-Author Paper March 2026

Hybrid Multi-Agent Conversational AI Architecture

Introduces a 10-agent modular AI framework with supervisor routing, competitive debate between agents, structured verification, expertise-aware personalization, and LLM-as-a-Judge evaluation for conversational AI systems. Achieves +11.9% accuracy over single-model baseline and 31.4% fewer unnecessary tool calls.

Multi-Agent Systems LLM Evaluation CrewAI Llama 3.3 Supervisor Routing Agentic AI

View Repository

Synopsis 2025

Cloud Removal for LISS-IV Satellite Imagery

Formal research synopsis covering the architecture, methodology, evaluation approach, and remote-sensing context for a physics-guided computer vision system for cloud removal in LISS-IV satellite imagery. Uses SAR-optical fusion and NDVI preservation, evaluated through PSNR and SSIM metrics.

Computer Vision Remote Sensing SAR Fusion PyTorch PSNR / SSIM NDVI

Open Write-up