I lead Spatial AI Interfaces Project/Lab at Amazon Lab126, where we invent new interfaces using multimodal sensors that can interpret the physical world for adaptive and ambient understanding.

Previously I led HCI Research and Product for Usefulsensors, a spinoff with the founding Team of Tensorflow (Google Brain) – building private/locally running AI agents on low-compute hardware using speech, LLMs, and TTS models. Discovering cognitive, bio/wellness markers from ambient human speech. I earlier co-founded and led Ducere, a wearable wellness tech startup developing a haptic I/O platform for the visually impaired and elderly for fall and gait analysis (exited 2018). And later Graviky Labs, an electrochemical carbon emissions CO2 to product venture

As a graduate research assistant at the MIT Media Lab, Massachusetts Institute of Technology I built AR gestural interfaces using transparent displays under Prof. Pattie Maes. At Hewlett-Packard Labs and Google I developed an open-source multi-projector stereoscopic (3D) multimodal gestural interface for CAD modeling (SketchUp).