Keras

Predict disease risks with AI-powered Django app offering role-based dashboards for patients, doctors, and admins.

Automate Adopt Me grinding with auto tasks, pet care, rewards collection, and 24/7 AFK farming for Roblox.

Improve LLM alignment with Selective Importance Sampling, reducing variance from off-policy tokens in RL post-training.

Unlock, analyze, and clone Dallas iButton keys with your Flipper Zero using Talos—a powerful tool for fast, accurate emulation and security testing.

Match property partners, prioritize by performance, and prepare WhatsApp outreach from your browser.

Train robotic systems with a large-scale synchronized vision, touch, and audio dataset for sliding contact tasks.

Execute algorithmic trades using a high-performance system built in Rust for speed, memory safety, and reliable backtesting.

Implement a round-robin task scheduler on the STM32F4 without an RTOS or HAL library using custom assembly startup and linker scripts.

Classify galaxy images into spiral, elliptical, or irregular types using a convolutional neural network and the Galaxy10 DECals dataset.

Comparative study of shallow vs deep CNN architectures on Fashion-MNIST — covers data preprocessing, model training, performance comparison, and error analysis with confusion matrices.

Convert Markdown into styled, selectable HTML code blocks directly in your browser.

Visualizer for neural network, deep learning and machine learning models

Run local LLMs with TurboQuant KV cache compression for 4-6x more context on the same GPU and an OpenAI-compatible API

Build AI research reports from a topic with Spring Boot, Spring AI, and Browserbase in one API call

Build and manage a local AI workspace that preserves your chat history, supports growth, and adapts to your personal needs.

Enhance your Adopt Me! gameplay with a Lua script that automates tasks, speeds up pet hatching, and unlocks hidden areas on Roblox.

ResU-Net for breast tumour segmentation on ultrasound images (0.9871 accuracy). Published in IJERCSE, Vol. 10 Issue 1, 2023.

🛌 Detect drowsiness and emotions in real-time using your phone camera with a FastAPI backend and React Native app.

🛠️ Detect casting defects using neural networks with this Python implementation, featuring MLPs and CNNs for effective image-based classification.