SYSTEM_READY // ARCHIVE_01

Engineering
Intelligent Systems.

Architecting the intersection of generative intelligence, autonomous agents, and large-scale defensive AI protocols.

01_SELECTED_WORK

PROJECT_COUNT: 6

ukti.ai

Voice Agent Platform designed for high-concurrency real-time interactions. Leverages low-latency speech-to-text synthesis and adaptive prompt injection for natural dialogue flow.

Python LangGraph vLLM WebRTC FastAPI GCP

DGIS — Indian Army

Specialized AI framework for the Indian Army. Automated document digitization and summarization on air-gapped, offline-only infrastructure.

PyTorch vLLM OCR GIS

Autonomous Desktop Control

A multimodal agent capable of navigating complex GUI environments. Employs vision-language models to interpret screen states and execute precise keyboard/mouse workflows.

LangGraph vLLM Computer Vision Custom Datasets
SYSTEM_VISUALIZATION_PENDING

Hospital Agentic Chatbot

Natural language chatbot for hospital service request routing via WhatsApp. Staff management dashboard with LangChain orchestration.

LangChain WhatsApp API Python

Automated Procurement Agent

AI-driven procurement automation for invoice and PO processing in Microsoft Teams with Adaptive Cards.

MS Bot Framework Adaptive Cards Teams

GenUI Extractor

Declarative GenUI Extractor that generates structured JSON UI trees from LLM tool call results.

Python JSON Schema
View on GitHub

02_RESEARCH_&_PUBLICATIONS

ACADEMIC_CREDENTIALS
Honorable Mention Sir C.V. Raman Award

CoLeafNet: A Dual-Track Deep Learning Network for Classification of Nutrient Deficiencies in Coffee Leaves

using DenseNet and Efficient Multi-Scale Feature Attention

Pioneered a dual-track convolutional neural network architecture for automated precision agriculture. Detecting nutrient deficiencies in coffee crops from leaf images with high accuracy.

Published in: Scopus-indexed Journal
Year: May 2025
Field Study IOP Science

Image-based Recognition of Environmental Microorganisms

using Deep Learning and CNNs

Research into automated identification of microbial structures using residual learning frameworks. Multi-class classification for bacteria, algae, fungi from microscopy datasets.

Published in: IOP Science Journal
Year: 2024