Amod Jog
About Me
I am a computer vision and medical imaging scientist based in Pune, India. I have significant experience in building AI/ML solutions in healthcare imaging, manufacturing, and ecommerce sectors. Presently I am a principal scientist at TeraRecon Inc, a ConcertAI company.
Previously, I was a lead scientist in the Industrial AI lab at GE Research Bengaluru in the computer vision team, where I developed AI/ML algorithms to improve scanning efficiency in MRI scanners and functional MRI processing. I also built signal processing algorithms for sensors for non-destructive evaluation of industrial parts.
Prior to that I was a computer vision scientist at IBM Watson Health where I built object detection and image classification algorithms using the (then) latest deep learning architectures for cancer detection in digital mammograms. I also did a couple of postdoctoral stints at Massachusetts General Hospital (MGH) and Johns Hopkins University (JHU) primarily working on developing image segmentation and image synthesis algorithms for brain MRI.
- Medical Image Analysis
- Computer Vision
- Natural Language Processing
PhD in Computer Science, 2016
Johns Hopkins University
MSE in Computer Science, 2011
Johns Hopkins University
BTech in Computer Science and Engineering, 2009
Indian Institute of Technology Bombay (IITB)
Experience
Responsibilities include:
- Developing computer vision and machine learning applications in healthcare–particularly processing cardiac and head and neck CT images.
Responsibilities include:
- Developing computer vision and machine learning applications in healthcare–particularly spine MRI and functional neuroimaging–and aviation–non-destructive evaluation signal processing.
Responsibilities include:
- Developing image classification and detection algorithms for screening mammography and chest X-rays
Projects:
- Developing image synthesis and image segmentation approaches to reduce the bias and variance in neuroimaging analysis results introduced due to MR scanner differences
- Human infant (age 0-2 years) brain MRI skullstripping application using convolutional neural networks
- Classification of healthy vs mild cognitively impaired (MCI) subjects using random forests on reaction time data
Projects:
- Estimation and analysis of strain during speech in the tongue muscles using diffusion and tagged MRI data
- Self super-resolution of thick slice 3D MRI scans
Projects:
- Designed machine learning algorithms that, in concurrence with MRI image formation physics, synthesize brain MRI images of a target modality using available MRI images.
Projects:
- Designed metrics and systems to classify skill levels and propose feedback to surgeons operating the da Vinci surgical robot using kinematic and stereo video data streams
Recent Posts
I was in Mumbai in our Powai campus, on 26th November 2008, preparing for my B.Tech project presentation. I was following the online news of the Pakistani terrorist attack on the Taj, Trident, CST, Chabad House at Nariman Point, Madam Cama Hospital, Leopold Cafe–pretty much all the places that make Mumbai Mumbai. I remember being slightly concerned and only mildly frustrated by it. I had been more or less desensitized to Islamic terrorist attacks in India and the consequent loss of innocent lives. There had been a rash of Islamist terrorist bomb attacks in India in the first decade of 2000s and somehow we seem to have had normalized them. In 2006, I was at home for summer break when multiple bombs exploded in the Western Railway suburban local train in Mumbai, killing 209(!) & injuring 700 and more. I used to take that local quite often from Dadar to Churchgate when going to South Mumbai. As I recall, even then my reaction was rather phlegmatic. “This shit just happens and we Indians just have to live with it” was the prevailing societal–and probably even my own–attitude.
