Ever since childhood, Chinmay Jog displayed a natural inclination towards mathematics and science, nurturing a strong desire to pursue a career in technology. True to his aspirations, he has successfully achieved that goal and now works as a Senior Machine Learning Engineer at Pangiam, a company based in Virginia state of the USA which offers innovative computer vision solutions and services for the travel and transportation industry. Born and brought up in Pune, Maharashtra, Chinmay opted for Electronics and Telecommunications as his major for engineering and joined MIT Pune, one of the top engineering colleges in Maharashtra. The signal and image processing coursework and projects excited him about computer vision and motivated him to pursue his Master’s degree from the University of Southern California.
“I got excited by the possibility of computers being able to perform human-level tasks while working on some image processing projects during my bachelor’s degree. Diving into computer vision and machine learning was a natural step to advance my knowledge, which led me to pursue my Master’s degree at the University of Southern California”, recalls Chinmay.
Eager to explore different fields in computer vision, Chinmay became a research assistant at the radiology lab of USC. He researched advanced image processing and machine learning techniques to tackle differentiating malignant tumors from benign ones using CT scan images.
His research was published in SPIE conference proceedings and further prepared him for a career in applied research and AI.
After graduating from USC, Chinmay joined Trueface, an early-stage AI startup focused on Face Recognition (FR), as one of the first research team members. While at Trueface, Chinmay worked on various computer vision projects using a subset of machine learning methods called deep learning.
“It was always my dream to work in a startup to get the maximum exposure to running a business while working in a technical capacity. I am glad I had the opportunity to work with such an amazing team at Trueface.”
Deep learning algorithms are notorious for requiring large volumes of data to work accurately, and the FR industry has always faced scrutiny over misuse, bias, and privacy concerns regarding biometric data collection. To develop ethical bias-free models without increasing costs, Chinmay researched non-data-intensive ways to eliminate bias from FR models. This research led to solutions that contributed to Trueface winning US Air Force grants for using FR for access control and a contract with a crypto-currency bank for identity verification, among other commercial agreements. Furthermore, Chinmay developed smaller and faster FR models for mobile devices. In addition to face recognition systems, Chinmay contributed to other computer vision applications like age detection from images, and liveness detection, which aims to identify attempts by unauthorized persons using photos and videos to bypass the face recognition system by impersonating someone else.
Trueface also submitted to the NIST face recognition vendor test (frvt), a global performance benchmark for FR models. Chinmay developed an FR model which ranked amongst the top fifteen in one category at the time of submission out of almost two hundred entries in the frvt challenge, performing equally well across all groups of races and gender. This performance played a crucial role in the acquisition of Trueface by Pangiam. At Pangiam, Chinmay has continued to work on face recognition model development in areas of performance and speed.
The Pangiam research team, Chinmay notes, has continued to build on the solid foundation of Trueface. Most recently, Pangiam’s FR solution has ranked first in the US Department of Homeland Security’s biometric rally 2022 in all categories. The DHS result and substantial commercial successes have cemented Pangiam’s name as one of the leaders in Face Recognition.
“The ML and AI space in India is booming, and I see massive potential for FR applications in India. Mobile payment ecosystems, targeted subsidies, and timekeeping systems are good examples of how businesses can use Face Recognition for identity verification and authentication to make the existing systems more secure”, Chinmay says.
In addition to hard work, Chinmay credits much of his success to the mentorship and guidance from his professors, managers, and peers over the years. He strongly believes in guiding and helping students and budding ML engineers, and he regularly participates in AI hackathons as a mentor and a judge. In the short term, Chinmay plans to continue developing applications using computer vision and machine learning to drive positive change. He also plans to continue serving as a peer reviewer for scholarly journals and mentoring and judging in AI hackathons to grow personally and professionally.
