NUS GRIP Run 4 participants. Led the team from ideation to initial partnerships. Among the top 18 teams selected to present on the final lift-off day.

Developed ANDON system for Whirlpool. Led the team to build prototypes for fridge quality inspection, forklift reverse collision alert, bottle inspection, ANPR, and under-vehicle inspection system.

Helped in developing CES award-winning smart-home assistant in Las Vegas 2018. Led ASR division, and assisted Computer Vision & NLP teams. Designed the architecture for the entire ASR suite.

Our team was among the top 12 finalists for India's biggest driver-less car challenge - 'Mahindra Spark The Rise'.Major contributions:✔ Real-time object detection on Indian Roads using Deep Learning for driverless cars.✔ Lane detection using Deep Learning and computer vision techniques.✔ Parsing and decoding data from CAN bus for controlling the car remotely.✔ Training models on newly collected data and producing desired results.✔ Designing Graphical User Interfaces on… Show more Our team was among the top 12 finalists for India's biggest driver-less car challenge - 'Mahindra Spark The Rise'.Major contributions:✔ Real-time object detection on Indian Roads using Deep Learning for driverless cars.✔ Lane detection using Deep Learning and computer vision techniques.✔ Parsing and decoding data from CAN bus for controlling the car remotely.✔ Training models on newly collected data and producing desired results.✔ Designing Graphical User Interfaces on Linux for: ⇨ controlling car-actuators and sensors remotely, ⇨ running deep learning framework 'YOLO' without using Linux Terminal, and ⇨ other tedious, time-consuming yet necessary data preparation tasks required for training models.Major contributions to other projects:✔ Hotword detection for a floor mopping robot.✔ Baby cry detector for a security camera project.✔ Geo-fencing with an alert system for security cameras.✔ Real-time object-tracking in crowded scenes.✔ Contributed to slow motion and panorama algorithms for the 'REOS' camera app by Cube26. Show less

✔ Worked on car detection using conventional machine learning techniques - HAAR, HOG, etc.✔ Immensely contributed to large dataset preparation.✔ Extensively researched about the problem of real time object detection on Indian Roads and presented deep learning as the solution over conventional computer vision techniques.