Karthik Govindappa is a Head of Data Science and Machine Learning @nyris GmbH at nyris GmbH. He possess expertise in deep learning, computer vision, image processing, robotics, opencv and 12 more skills. He is proficient in German and English. Colleagues describe him as "Karthik worked with me as Machine Learning engineer in building BuddyGuard, a revolutionary home security system (www.buddyguard.io).
Karthik is one of the early employees, which had to take big responsibilities from the early days, in order to help with the demands of the project. Flare is using 5 ML algorithms - face detection, face recognition, object detection, sound recognition (speech, glass breaking, door knocking, fire alarm, steps, etc.), speaker recognition, speech recognition.
Karthik started as a Machine Learning intern focused on image processing algorithms, but soon he would handle very complex tasks on all ML fronts. Karthik he quickly catched up with the work of several experienced ML engineers on the fields of FaceRecognition, Object Detection and Sound recognition. Not only he catched up fast with all our algorithms above, but he quickly started to add extra value to all our work. He would research and apply various compression technologies to all our CNNs and develop/implement computationally efficient algorithms based on OpenCL in order to take advantage of our graphics embedded card.
Karthik would be the man to take any challenge, focus on it completely, work on it in silence and don’t let it out of his hands, until he has something great to show his colleagues. He would take great joy in accomplishing a task over the expectations, even if that meant, staying late hours in the office, just to try new approaches learned on different papers he was consulting. His energy and dedication were amazing, and definitely has inspired many of us.
From a technical perspective, he was responsible of building/maintaining artificial intelligence algorithms to power the intelligence of IoT Security devices, in this case Flare. Among the challenges, that he successfully worked on, I can enumerate :
* Improving and testing different deep learning architectures for sound recognition, face recognition, object detection/recognition
* Building and maintaining large face/objects/sound datasets
* Building and Improving the training pipelines of our ML models
* Research and implement compression technologies to various CNNs
* Synchronise and optimize parallel analysis of audio and video processing on a limited IoT device
* Develop prototypes for new algorithms we were trying - SSD, Yolo, etc.
* Develop tools to deploy ML models and firmware on many devices, teached other people to use these tools
* Create machine learning toolkits (buddycv/buddysound) libraries to be used with main firmware
etc.
Karthik he was really excited to adopt new trends in ML and other fields, even if they would take a great deal of learning. He was always researching into new papers, follow AI conferences, keep everyone from the team informed about his discoveries. Many times we’ve got surprised about his desire to enlarge his knowledge on all fronts, even if that was not his main speciality - he was providing solutions related to hardware, solutions for faster p"
Listed skills include Deep Learning, Computer Vision, Image Processing, Robotics, and 13 others.