Welcome to my universe
Computer Applications undergraduate engineering real-time 3D systems, biological Graph Neural Networks, and low-latency IoT firmware — bridging machine intelligence and physical silicon.
FruitFly Connectome
GNN & Hybrid Chess
Pantheon Gateway
Universal AI Studio
Stellar Hands & IoT
500K Particle Engine
I'm a Computer Applications undergraduate specializing in Machine Learning, Low-Level Systems, and Real-Time Graphics. I engineer high-throughput AI serving gateways, formulate Graph Attention Networks (GAT) on biological connectomes with symmetric INT8 quantization, and craft bare-metal C++ firmware for ESP32/ESP8266 silicon.
Whether optimizing FreeRTOS ring buffers for 400 kHz OLED radar telemetry, instancing 500,000 WebGL particles to webcam hand landmarks at 60 FPS, or accelerating diffusion models on Apple Silicon Metal Performance Shaders (MPS), I build with zero compromise on efficiency and latency.
Biological Graph Neural Network (GNN) and hybrid chess engine. Adapted a 256-neuron Drosophila connectome topology into a 3-layer Graph Attention Network (GAT) in PyG, quantized to symmetric INT8 (7.12 MB) with <0.6% precision loss, fused with Stockfish 19 NNUE and an interactive 3D WebGL telemetry simulation.
Universal multi-model AI gateway and serving studio. Asynchronous ASGI routing across local models (Ollama/llama.cpp) and cloud APIs with real-time TTFT and TPS latency telemetry, plus local text-to-image (SDXL Turbo) accelerated on Apple Silicon MPS.
Real-time gestural computing system running 500,000 GPU-instanced particles entirely in-browser. Tracks webcam hand landmarks via MediaPipe to pull, repel, freeze, or shockwave cosmic particle formations across ten dynamic topologies at 60 FPS.
Passive 802.11 RF mapping and spatial presence monitor on ESP32 silicon in promiscuous mode. Captures management frames and fits raw RSSI telemetry into a log-distance path loss attenuation model to estimate proximity without network pairing, driving an I2C OLED radar display.
Real-time sign language interpreter — recognises dual-hand ASL gestures via webcam tracking 21 3D landmarks per hand to identify 30+ signs with 0ms server latency, using angular and Euclidean joint vector classification with temporal hysteresis filtering.
Palm-sized interactive desktop companion robot built on ESP8266 silicon with an SSD1306 OLED display. Features a non-blocking Finite State Machine (FSM) coordinating expressive bitmap eye animations, behavioral state changes, and capacitive touch sensors handled via hardware ISRs.
A technical skillset bridging biological Graph Neural Networks, universal AI serving architectures, real-time 3D WebGL graphics, and bare-metal embedded firmware.
Local AI & Silicon Acceleration — runs, serves, and accelerates LLMs and diffusion models on personal hardware with Apple Silicon MPS. Zero cloud lock-in, zero cost, full control.
Biological GNN & Connectome Modeling — trained Graph Attention Networks on 256-neuron Drosophila topology with symmetric INT8 quantization (<0.6% deviation) fused with Stockfish NNUE.
Hardware + Software Co-Design — bridges the gap between low-level embedded C++ firmware on ESP32/ESP8266 and real-time browser-based WebGL graphics.
Scale-Minded Systems Builder — optimized systems from 500,000-particle GPU instancing to 400 kHz I2C sensor telemetry loops. Latency and memory efficiency are first-class defaults.
Bachelor of Computer Applications (BCA), Gaya College, Magadh University (2025–2028). Certified in IoT & C/C++ Systems Programming (Udemy, 2026).
Artificial Intelligence Fundamentals — recognized for professional excellence in core machine learning and AI concepts.
Craft Precise Prompts for AI Models — practical training in structuring effective system prompts and conversational architectures.
Cybersecurity Job Simulation — completed practical industry tasks in cybersecurity, defense fundamentals, and enterprise threat analysis workflows.
Data Science Job Simulation — completed end-to-end industry tasks: business framing, EDA, data cleaning, feature engineering, modeling, and strategic insights.
TCS iON Career Edge – Young Professional — covering communication, workplace thinking, presentation, resume writing, interview skills, business etiquette, and AI overview.