The true impact of AI in solving real world problems will be realised only by creating more transparent and useful models that can bridge the gap between technology and people who benefit from it !
Prateek Singh, our next pathbreaker, Senior Technical Lead at Mercedes-Benz R&D (India), leads a team focused on developing AI-driven solutions for real-time sensor data processing.
Prateek talks to Shyam Krishnamurthy from The Interview Portal about his PhD at IIT Roorkee, where he developed an explainable deep learning model to classify arrhythmias in ECG data of patients, and subsequently leveraging his background in physiological signal processing and biomedical AI, with a focus on automotive safety.
For students, persistence and a willingness to adapt are essential, especially in fields like AI and technology, where things evolve quickly.
Prateek, what were your initial years like?
I grew up in a small city in India and, like many kids there, I was an average student. Opportunities and career paths weren’t always visible to us, and in our town, there were limited, familiar paths to choose from after school. If you studied math, like I did, engineering was the expected next step. For biology students, it was MBBS; for commerce, BBA or B.Com. Since I was in the math stream, engineering was a natural choice.
My parents supported this decision, and like many others from similar backgrounds, I set my sights on the IITs. Aiming to get into an IIT meant moving to Kota for coaching, which marked the beginning of my real struggles. Back in my CBSE school, teachers often focused on board exam content, while IIT preparation required a deep understanding of fundamentals and the ability to apply concepts in new ways. The IIT-JEE exam was designed to test true learning, not just rote memorization, and adapting to this change was a challenge. I quickly realized that my previous education had been more about passing exams than actually learning the material in a way that could be applied.
That one year in Kota was intense. Being away from my family for the first time, I had to learn to manage everything on my own—my food, accommodation, and my studies. Despite the hard work, cracking IIT-JEE was tough; back then, only seven IITs existed with limited seats, and I couldn’t make the cut. Instead, I joined a private engineering college, hoping for hands-on experience with electronics and real engineering projects.
However, college life was another reality check. The curriculum was mostly theoretical, with an emphasis on passing exams. Hands-on learning was limited, and I felt as if each year was a repeat of the last—study, take exams, pass. By the end of my bachelor’s degree, I felt I had only scratched the surface of what engineering could be.
But I knew I wanted more. Rather than accepting a job in the IT sector, I decided to pursue a master’s degree, believing that with a deeper understanding, I could find the hands-on experience that I was seeking.
What did you do for graduation/post graduation?
After completing my bachelor’s degree (BTech, Electronics & Instrumentation), I wanted a deeper, more practical experience in engineering. I took the GATE exam, and with a good score, secured admission to MTech in Biomedical Instrumentation at the College of Engineering Pune (COEP), one of India’s respected government colleges. This was a refreshing change. COEP had an environment where students were actively involved in real projects, and I finally found the hands-on learning experience I had been searching for.
What made you choose such an offbeat, unconventional and uncommon career in Biomedical AI?
My career path was driven by a blend of academic curiosity and professional exposure.
My time at COEP (MTech) helped me build a solid foundation, especially through my work on ECG signals, which sparked my interest in healthcare technology. This interest continued to grow, and I decided to pursue a PhD at IIT Roorkee, one of India’s top institutes, where I specialized in ECG signal processing using deep learning. Here, I had the opportunity to work with Dr. Devi Shetty, a leading cardiologist in India, which changed my perspective of healthcare technology and inspired me to explore explainable AI techniques for medical applications.
During my PhD at IIT Roorkee, I became fascinated with the challenges cardiologists face when interpreting arrhythmias (An arrhythmia is an irregular heartbeat, meaning the heart beats too quickly, too slowly, or with an irregular rhythm) in ECG data. This focus on ECG signal processing was inspired by real-world problems in healthcare, particularly in collaboration with Dr. Devi Shetty, a prominent cardiologist. Dr. Shetty’s emphasis on the need for interpretability in AI models gave me insights into the limitations of “black-box” AI solutions in healthcare and sparked my interest in explainable AI.
These experiences reinforced my goal of applying AI and machine learning in medical applications, aiming to build solutions that not only deliver accurate predictions but are also transparent and usable for healthcare professionals. Working on this project gave me a deeper understanding of how technology could solve pressing healthcare issues, guiding me toward a career in biomedical AI and signal processing.
How did you make a transition to a new career? Tell us about your career path
My career has been shaped by a series of well-planned steps and key decisions aligned with my passion for biomedical AI. After completing my bachelor’s in Electronics and Instrumentation, I realized I wanted more than just theoretical knowledge. This led me to pursue a Master’s at COEP, where I gained hands-on experience in biomedical instrumentation, especially through my work on ECG signals. This work deepened my interest in healthcare technology, motivating me to further my research through a PhD at IIT Roorkee.
During my PhD, I was offered a consulting role at Elastic Care, a Canadian medical device startup focused on developing Holter monitors. This opportunity allowed me to apply my academic research to real-world applications. At Elastic Care, I worked on developing various algorithms, including heart rate estimation, arrhythmia detection, and ECG denoising. These algorithms were integral to enhancing the capabilities of their Holter monitors. Working on projects directly aligned with my PhD research gave me valuable insights into the practical steps needed to bring medical technologies to market, particularly in navigating the stringent requirements for safety and effectiveness, including FDA standards and medical device testing.
These experiences not only enhanced my skills but also led to multiple job offers from academia and industry before I completed my PhD. Due to family commitments and the opportunity to work on projects that aligned with my expertise, I accepted a role at Mercedes-Benz R&D India. Here, I lead a team working on AI and real-time signal processing, a position that combines my academic background with industry experience, allowing me to make impactful contributions in both automotive safety and biomedical applications.
How did you get your first break?
My first major break came during my PhD when I was offered a consulting role with Elastic Care, a healthcare startup in Canada. This was a direct hire for an overseas position, which is quite rare, especially since I had not yet completed my PhD. I got this opportunity through networking and by demonstrating the effectiveness of my algorithms. Elastic Care found my work compelling, particularly because my PhD research was perfectly aligned with their focus on ECG signal processing and medical device development.
The role was remote, which allowed me to continue my PhD while working with Elastic Care. This opportunity allowed me to bridge my academic research with real-world applications, especially in the development and testing of medical devices. Working with Elastic Care gave me hands-on experience in meeting FDA standards and understanding the regulatory landscape, which was incredibly valuable and closely aligned with my PhD work. This role was a pivotal point in my career, as it not only enriched my practical knowledge but also positioned me for future roles in both academia and industry.
What were some of the challenges you faced? How did you address them?
Challenge 1: Adapting from a small-town educational background to the rigorous demands of IIT preparation was a significant challenge. The transition required me to not only catch up academically but also adjust to a new, independent lifestyle in Kota, far from home. I learned time management and self-reliance through this experience, which laid the foundation for my future endeavours.
Challenge 2: During my PhD, I faced the difficulty of developing explainable AI techniques for ECG analysis. This required intensive research into novel methods that could make AI decisions interpretable for cardiologists. Collaborating with experienced professionals, like Dr. Devi Shetty, and immersing myself in the latest AI research helped me overcome this challenge, ultimately allowing me to create more transparent and useful models.
Challenge 3: Balancing my consulting role at Elastic Care with my PhD research was also challenging. Both required significant dedication and focus. However, my passion for the field kept me motivated, and I learned to prioritize and manage my time efficiently to meet the demands of both roles.
Where do you work now?
I currently work as a Senior Technical Lead at Mercedes-Benz R&D India. In this role, I lead a team focused on developing AI-driven solutions for real-time sensor data processing. My projects align closely with my background in physiological signal processing and biomedical AI, but with a focus on automotive safety. We work with multi-sensor systems, including cameras and Lidar, to enhance environmental awareness and improve vehicle safety features.
What’s a typical day like?
A typical day involves managing project roadmaps, setting technical priorities, and engaging hands-on with machine learning tasks to support my team. This role is fulfilling as it allows me to combine my expertise in AI with the opportunity to lead innovative projects that have real-world impact, not only in the automotive sector but also potentially in broader applications of safety and signal processing.
How does your work benefit society?
My work contributes to improving automotive safety by enhancing a vehicle’s environmental awareness through advanced sensor data processing. This can potentially reduce accidents and improve road safety, benefiting society at large.
Previously, my work with Elastic Care focused on developing medical devices and algorithms for arrhythmia detection, which had direct applications in healthcare. By enabling early detection of heart conditions, these technologies support timely medical intervention, contributing to better patient outcomes and healthcare quality. My career has been driven by the goal of applying technology to solve critical problems, whether in healthcare or automotive safety, ultimately making a positive impact on people’s lives.
Tell us an example of a specific memorable work you did that is very close to you!
One of my most memorable projects was developing a deep learning model to classify arrhythmias in ECG data. Working with Dr. Devi Shetty, a renowned cardiologist, gave me invaluable insights into the medical field’s needs and the importance of model transparency. Cardiologists require interpretability, so they can trust and understand AI decisions in critical diagnoses. This led me to incorporate explainable AI techniques, allowing the model to provide insights that clinicians could rely on.
This project holds a special place in my career as it combines cutting-edge AI with real healthcare applications, impacting both the technology and the people who benefit from it.
Your advice to students based on your experience?
My advice is to stay curious, resilient, and open to continuous learning. Don’t limit yourself to the curriculum; explore projects and internships that let you apply what you learn in real-world situations. Embrace challenges—they often lead to the most meaningful growth.
Networking and seeking mentors can also provide valuable insights and guidance. Ask questions, collaborate with others, and learn from their experiences. Persistence and a willingness to adapt are essential, especially in fields like AI and technology, where things evolve quickly. Focus on building a strong foundation, and the opportunities will follow.
Future Plans?
I plan to continue advancing AI applications in healthcare and automotive safety, leveraging my experience to create solutions that are impactful and accessible. Specifically, I aim to focus on making AI models more interpretable and trustworthy, especially in sensitive fields like healthcare. This focus on explainability will help bridge the gap between advanced technology and practical, user-friendly applications.
I’m excited to work on projects that bring AI closer to everyday life in a safe and reliable way, whether through enhancing automotive safety features or developing diagnostic tools in healthcare.