About Krushna Parmar

Krushna Parmar is an AI/ML engineer with a deep curiosity for understanding the inner workings of technology, from how models learn to how memory moves on a GPU. This passion drives the creation of small, sharp, and often open-source tools designed to make complex concepts clear and accessible. Currently pursuing an M.Tech in Machine Learning, Krushna focuses on building practical solutions and exploring the fundamentals of computing.

• HuskML: A visual, browser-based neural network designer that allows users to drag layers, visualize the computational graph, and export working Keras/PyTorch code. This project aims to democratize neural network design with a no-code approach.

• Exodus: An AI-powered meeting monitor that incorporates Whisper for transcription, Langflow for AI summaries, and facial recognition capabilities.

• Adrishyam: A computer-vision package available on PyPI for haze removal, utilizing the Dark Channel Prior method, complete with a live demo.

• Eco-Sattva: A MERN-stack application designed to track carbon footprints, integrating ML for sustainability metrics.

• AR Campus Navigation: An ARCore-based indoor campus navigation system that functions without GPS, relying on geolocation and QR code recognition.

• Chat App IoT: A local host chat application utilizing ESP32 and Bluetooth BLE with a custom user interface.

Krushna also shares insights and learnings through a public notebook, covering topics in computer vision, ML, and various technical rabbit holes. The portfolio emphasizes a commitment to open-source development and a hands-on approach to understanding and building technology.