AI Engineer & Full Stack Developer actively researching, learning, and engineering state-of-the-art AI systems. Creator of the fine-tuned Tejapriyan-8B open-source language model and Teja Priyan AI multimodal intelligence platform. Dedicated to solving complex problems through clean code, deep learning, and continuous AI innovation.
12+
Projects Built
8B
Custom LLM Parameters
100%
Dedication to AI
Passionate AI Engineer and continuous learner specializing in fine-tuning open-source LLMs, multimodal AI architecture, and computer vision systems. Creator of Teja Priyan AI (multimodal intelligence platform) and the Tejapriyan-8B language model trained with SFT and GRPO for verifiable SQL reasoning.
Completed a Computer Science degree in May 2025 and comprehensive enterprise Java Full Stack development training in 2025–2026. Actively advancing skills in deep learning, autonomous cognitive routing, reinforcement learning with execution sandboxes, and production deployment.
Open-Source AI Research & Independent Systems | 2024 – Present
Trained and fine-tuned the Tejapriyan-8B language model using Supervised Fine-Tuning (SFT) for identity alignment and GRPO reinforcement learning with verifiable SQL execution rewards. Architected Teja Priyan AI, an enterprise multimodal intelligence platform with sub-second token streaming, live code and game preview sandbox, vision reasoning, and autonomous multi-tier cognitive routing.
Professional Training Program | September 2025 · 6 Months
Completed comprehensive enterprise training in Java, Spring Boot microservices, database architecture, RESTful API design, and modern scalable application development.
University Graduation | Completed May 2025
Graduated with a strong foundation in data structures, algorithms, software engineering principles, distributed systems, and computer architecture.
Academic & Research Projects | 2023 – 2024
Implemented real-time deep learning systems including helmet detection, license plate recognition, and smart traffic monitoring using Python, OpenCV, and YOLO models.
AI & LLM Fine-Tuning
Multimodal Systems
Computer Vision
Full Stack Engineering
Explore key projects across Artificial Intelligence, Large Language Models, Computer Vision, and Full-Stack Web.
Website Architecture & Security Engine
Recursive website intelligence platform dissecting full-site structure, security posture, DNS headers, tech stacks, and internal networks in real time.
Encrypted Peer-to-Peer Data & Clipboard Bridge
Zero-install cross-device AirDrop alternative connecting phones and PCs via WebRTC data channels, animated QR signaling, and end-to-end encryption.
Multimodal AI Workspace & Reasoning
Real-time streaming multimodal AI workspace with <0.4s latency, 4 adaptive depth tiers, native diagram reasoning, and live interactive code & game sandbox.
Fine-Tuned LLM · Verifiable SQL Reasoning
Fine-tuned 8B parameter model with SFT and GRPO execution rewards (57.7% on Spider benchmark). Runs offline with Ollama, in terminal via npx tejapriyan, and SQLite Wasm.
Kinetic Typography & Physics Studio
Interactive canvas laboratory powered by Matter.js rigid-body dynamics, procedural audio synthesis, and real-time interactive typography.
Audio-Reactive DAW & Synthesizer
Browser DAW converting microphone audio and music tracks into real-time reactive canvas visualizations and customizable frequency spectrums.
Pixel Reveal Studio & Video Generator
In-browser pixel reveal studio transforming images into dynamic particle animations and high-framerate reveal videos directly on client device.
Interactive Mystery & Escape Maze
Atmospheric browser puzzle game where the web interface falls apart into 8 hidden glitch chambers with terminal riddles and 7 unique endings.
AI & Computer Vision Compliance
Real-time computer vision system using convolutional deep learning models to identify motorcycle riders without helmets with 95%+ accuracy.
IoT & AI Smart City Platform
Real-time traffic density analysis, congestion forecasting, and violation detection combining YOLO object detection with IoT infrastructure.
Immersive Spatial Storytelling
Scroll-driven 3D web experience built with Three.js, React Three Fiber, Next.js, and Framer Motion with real-time particle and lighting physics.
Interactive Web Arcade
Curated arcade of browser-based mini-games highlighting dynamic physics, keyboard mechanics, and responsive HTML5 Canvas rendering.
I'm always interested in hearing about new projects and opportunities. Feel free to reach out through any of the channels below!