devashish-gupta
Deeply passionate about robotics, signal/image processing, radio astronomy and fluid mechanics.
Georgia TechAtlanta, GA
Pinned Repositories
devashish-gupta
Geode
A zero-shot geospatial question answering agent with precise spatiotemporal retrieval and explicit reasoning capabilities!
Instruct-Nav
A multimodal model for language-guided socially compliant robot navigation.
Pointy
Comparison of point cloud classification performance using 2D projection based and transformer based methods. Also explores adversarial attacks on these classes of models. Specifically, one-point and noise attacks.
Registr
A personal web-app for real-time course statistics monitoring and proactive registration planning. Designed for Georgia Tech students, enabling quick visual comprehension and inference-based decision making.
Tranquil
Tranquil is a breathtaking Algorithmic Reverb Plugin for all the major digital audio workstations (DAWs), meticulously crafted with an architecture that is adept in building prolific soundscapes and simulating small rooms with the same fidelity. Available as a 64-bit VST3 binary for Windows.
devashish-gupta's Repositories
devashish-gupta/Tranquil
Tranquil is a breathtaking Algorithmic Reverb Plugin for all the major digital audio workstations (DAWs), meticulously crafted with an architecture that is adept in building prolific soundscapes and simulating small rooms with the same fidelity. Available as a 64-bit VST3 binary for Windows.
devashish-gupta/Registr
A personal web-app for real-time course statistics monitoring and proactive registration planning. Designed for Georgia Tech students, enabling quick visual comprehension and inference-based decision making.
devashish-gupta/devashish-gupta
devashish-gupta/Geode
A zero-shot geospatial question answering agent with precise spatiotemporal retrieval and explicit reasoning capabilities!
devashish-gupta/Instruct-Nav
A multimodal model for language-guided socially compliant robot navigation.
devashish-gupta/Pointy
Comparison of point cloud classification performance using 2D projection based and transformer based methods. Also explores adversarial attacks on these classes of models. Specifically, one-point and noise attacks.