Building something groundbreaking to bring AI to life.

Founded by Andrew Ng, AI Link is a community for entrepreneurs and technologists working to build great AI companies together.

Following its acquisition by Google, Phiar has spearheaded the development and integration of edge-based perception and advanced driver experiences into vehicles, as part of Geo Auto within Google Maps.

Overseeing technology development, R&D, and product vision and strategy; growing and recruiting engineering and product teams; evangelist of AR and AI for smart mobility in the automotive space.

Harvard i-Lab '17, and YC S18.Phiar is an ultra-lightweight AI company, building the world's most advanced Ultra-Lightweight Spatial AI Engine to enable every vehicle to see its surroundings, running real-time at the edge, using just commodity mobile-grade processors and existing automotive computing platforms with a single camera.Focused first on powering our Intelligent AR Navigation Platform, as the anchoring/foundational use-case to transform driving experience and safety, and to ultimately enable the next-gen in-vehicle experiences and connected cars services.

Provided expertise in deep learning techniques and modeling advice to the modeling team on the development of human neural prosthesis.Designed and developed convolutional and recurrent neural network models to train and predict neuronal firing rate data collected from rodents and humans using Keras.

www.paintapic.comDesigned and developed an industry-first core image processing and computer vision engine that converts users' uploaded photos into paint-by-number templates. This engine operates in real time on the web service, providing users with live previews of their finished paint-by-number products.

Research area: Computer vision, machine learning, and brain-inspired model of visual features with deep learning. Created the first biologically-consistent CNNs for primate visual pathways and developed PGP, a novel clustering algorithm achieving state-of-the-art results in video segmentation. Advanced visual perception research through the CCF model for predicting human object recognition and visual clutter analysis.

Assisted during lecture, held office hours and review sessions for the following courses: Programming in Java, Analysis of Algorithms, and Computational Biology.

Developed and experimented with RNN and CNN models for text-to-image mapping, age estimation, and face detection using Caffe and Torch7, and contributed to the Caffe open-source community.

Researched and implemented various locality-sensitive hashing (LSH) schemes in C++ to optimize data de-duplication efficiency for the Whitewater cloud gateway product.

Conducted research focused on developing automated algorithms for 3D MRI brain tumor segmentation. Assisted in teaching Data Structures & Algorithms and Intro to Computer Graphics, including conducting office hours and review sessions.

Developed innovative 3D neuroimage processing methods for Alzheimer’s Disease diagnosis, involving data sorting, validation, and mining with machine learning and pattern recognition.

Developed mathematical and computational models for visual motion analysis, including a biologically inspired motion-sensitive system and a dual-Gaussian model to study human visual cortex interactions with optic flow. Designed and implemented cognitive neuroscience research software for visual navigation and memory impairment studies using Quake 3 engine code in C.