Research archive · 6 builds

Systems, not just models.

End-to-end AI/ML work spanning simulation, evaluation, data pipelines, interfaces, and measurable outcomes. Open a case study for the engineering details.

Research statusActively building

Core domainApplied AI systems

Evidence standardGrounded + evaluated

DeliveryEnd-to-end

6 of 6 projects

DERGuardian — Cyber-Physical Anomaly Detection

Active

Research pipeline for anomaly detection in Distributed Energy Resource (DER) systems: simulation-grounded telemetry, synthetic cyberattack scenarios, and multi-detector benchmarking.

PythonOpenDSSPyTorch+2

AI Expense Tracker with RAG-Enhanced Financial Reasoning

Done

Full-stack AI system converting natural language queries into SQL-grounded, LLM-explained financial insights.

PythonFastAPISQLite+6

Forest Fire Severity Detection (IEEE Publication)

Done

End-to-end computer vision pipeline using PyTorch and OpenCV to classify wildfire severity. Published on IEEE Xplore, SETCOM 2025.

PyTorchOpenCVComputer Vision+1

Tiny LLM for Anomaly Pattern Learning (Experimental Extension)

Active

Evaluating lightweight open-source LLMs for real-time anomaly classification from structured telemetry, explored as part of DERGuardian.

Lightweight LLMsTransformersPrompt Engineering+1

GearCom Smart Helmet

Done

IoT-based safety system for accident detection using embedded sensors with signal processing and real-time alert logic.

IoTSignal ProcessingEmbedded Systems+1

Navarang Engineering Works Website

Done

Built and deployed a company website to improve customer discovery and digital presence for a manufacturing business.

Web DevelopmentDeploymentBranding