Hi, I'm Rudransh, a 10th grader passionate about software engineering. I love building projects—from AI models and LLMs to electronics, automation tools, and cybersecurity experiments. Many projects here started as ways to solve real problems or explore what's possible. I've worked on custom hardware, trading bots, web apps, and tools to simplify AI workflows. This portfolio showcases some of my experiments, both complete and in progress.
Beyond tech, I'm into badminton, hiking, and cycling in the hills of Uttarakhand.
Below are details on some projects I've developed.
A specialized fine-tune of the Phi-3 model, optimized to enhance its capabilities in logical reasoning and problem-solving tasks. Developed using Python, this project focuses on improving model performance through precise data formatting and tuning. You can find it on Hugging Face: here .
A fine-tuned iteration of the TinyLlama model, engineered to deliver superior performance on Natural Language Processing (NLP) tasks. This project demonstrates efficient model optimization and dataset structuring. You can find it on Hugging Face: here .
Built this during a break for fun in Python. It includes a GUI for LoRA fine-tuning models, a simple hyperparameter tuner, VRAM optimizer, and model merger. Available on PyPI: here .
Dunebench is a lightweight, local benchmarking tool for GGUF models. It allows you to evaluate Large Language Models (LLMs) across a variety of domains—including logic, coding, math, and common sense—using llama-cpp-python. here .
SPEK is a simple Python library for real-time object detection using YOLOv8 and OpenCV. It supports webcam, video, or static image detection and is designed for ease of use. Available on PyPI: here .
IDR is a lightweight Python library to automatically enhance dark or overly exposed images. It automatically adjusts brightness and contrast. Available on PyPI: here .
An early version of gguf2gui.
Fascinated by AI and LLMs, I explored the efficient GGUF format and built this simple Python GUI tool to convert safetensors models to GGUF. It integrates with ANA too. Find it on PyPI.
A simple Python library for easy keyboard event detection—great for custom shortcuts or input-heavy apps like games. On PyPI.
Osinto locates and retrieves free online books, movies, and OSINT data to boost access to educational/entertainment resources.
Content Discovery: Advanced search to catalog free content from sources.
User-Friendly Interface: Easy CLI navigation.
Open-Source Intelligence: Provides valuable info without copyright issues.
Automated vulnerability assessment for WiFi networks—detects encryption issues, weak passwords, and misconfigs.
Network Packet Analysis: Recon for SSID, signal strength, and devices.
Exploitation modules to test vulnerabilities.
Password Security Scanner: Identifies weak passwords and shows Google/WiFi creds in plain text for awareness.
Malware Detection and Removal: Handles keyloggers and injected threats.
Security Enhancement: Emphasizes wireless network protection.
User-Friendly Interface: CLI with ASCII art.
FPS game with 30+ levels and responsive controls for precise gameplay.
Thrilling Scratch space game: Pilot through asteroids, dodge enemies, blast invaders. Upgrade ship and use special attacks.
Note: Made in 6th grade with my younger brother—early fun project!
This is just a selection of my work over the years. I have more personal projects in AI and automation.
