AI & LLM Applications
LLM-powered products, chat interfaces and applied artificial intelligence.
I build AI systems and agents
I build intelligent software systems across AI, C++, automation, networking, trading infrastructure and modern web applications.
My work ranges from AI-powered learning systems and RAG pipelines to low-latency C++ engines, remote-desktop infrastructure, synthetic trading platforms, computer-vision applications and developer tooling.
I enjoy working close to the system — understanding not only how an application looks, but how its AI, runtime, networking, data, security and performance layers work together.
Production-oriented AI and systems software — pairing intelligent agents with the fast, reliable, efficient infrastructure underneath them.
LLM-powered products, chat interfaces and applied artificial intelligence.
Retrieval pipelines, grounded chat, memory and agent tooling.
Low-latency engines, parsing, audio and desktop systems.
Detection, classification and image-understanding experiments.
Windows tooling, C# apps and Android projects in Kotlin and Java.
Remote desktop, screen streaming and device-to-device links.
Synthetic markets, trading agents and risk controls.
Wallet interfaces and token-based experiments.
React, TypeScript, Laravel and static sites on GitHub Pages.
CLIs, API checkers, PowerShell scripts and build utilities.
An AI-powered learning workspace combining document intelligence, RAG, knowledge graphs, AI tutoring, research and learning analytics. Processes real PDF, DOCX, Markdown, TXT and HTML sources into a grounded study system with quizzes, flashcards and runtime observability.
A low-latency, LAN-first remote-desktop system that turns an Android phone into a secure controller for a Windows PC — live screen streaming, touchpad, mouse gestures, keyboard input and bidirectional file transfer over LAN or hotspot, without relying on cloud infrastructure.
High-performance C++ trading-agent infrastructure focused on low-latency market interaction, automation and extensible trading logic — the newest direction in systems-level financial engineering.
A full-stack synthetic trading platform combining real-time market visualization with candlestick charts, automated trading bots, portfolio management and risk controls across multiple synthetic markets.
High-performance C++ document-processing pipeline designed for fast JSON parsing, transformation and PDF generation — a close-to-the-metal exercise in parsing and systems performance.
An AI-powered screen-understanding system that interprets visual desktop context for computer interaction — part of a broader exploration into desktop agents and screen-aware software.
Memory Twin AI — AMD Developer Hackathon 2026 project exploring semantic memory storage and retrieval: an AI-powered digital memory twin that stores personal memories and retrieves the relevant ones via embeddings.
Real-time C++ audio-processing experiments focused on improving speech clarity and reducing environmental noise during meetings.
Computer-vision system for identifying plant leaf diseases using deep learning, trained and run on AMD cloud GPUs with TensorFlow — from dataset to calibrated predictions.
Web3 wallet interface experiment exploring token-based applications, blockchain interaction and asset display — shipped as a live GitHub Pages app.
Early public work — synthetic markets, trading interfaces and a first API experiment.
A move toward C++, computer vision, machine learning and networking — plus the first web builds.
The shift into AI agents, multimedia, developer tooling and portfolio systems.
AI + C++ + networking + trading + desktop automation at once — the strongest output yet.
Education
Annamalai University