Tell me a little about your professional innings at Qualcomm

I have been at Qualcomm for 27 years - it's the only company I've worked for since grad school! Qualcomm too was a startup, in a sense. Founded in 1985 by university professors, the DNA of the company is research-based design and developing foundational technologies for cutting-edge communication. The company has retained its focus on R&D and is ubiquitous today for its mobile connectivity technologies. I have worked in wireless infrastructure, chipset customer engineering and for the past five years, with the patent licensing team. I was also involved with setting up the R&D centre in Hyderabad, and have been involved with India R&D since 2004.  

Qualcomm is poised to make AI ubiquitous, esp in areas like IoT, automotive, cloud and mobile phones. Can you elaborate on the company's plans on diversifying AI in these areas?

Before I delve into AI, let me give you some context into our origins in R&D and setting the global standards for quality communications. After designing CDMA for 2G, we had the early insight that the next interface via air would be data-driven. People found this to be implausible at the time, since SMS and GSM were adept at delivering messages and calls. Building data-based communication networks can be compared to building roadways - city roads are built to handle slow moving but heavy traffic, while freeways or highways need to manage vehicular movement at high speeds - similarly, low latency networks are ideal for low-resource settings, while high speed networks are better suited for densely populated pockets. This is the level of flexibility networks can provide, as with the progression and rapid adoption from CDMA to 5G. What's distinctive about data-driven communication is the sheer volume and density of data collected. And this is where AI on the device or AI on the edge is a game-changer. Leading cloud companies like Google, Amazon and Microsoft approach AI from the perspective of deep learning with large but context-specific datasets and depend on GPUs. Qualcomm's native advantage with powerful, battery-operated, handheld devices has brought AI to one's palm and on-the-go. Today, Qualcomm Snapdragon processors are present in almost every mobile phone, with immense speeds and compute power, capable of driving a range of functions. Nearly 150 AI programmes are already running on these phones. And with the advent of 5G, on-device AI capabilities are only going to grow further. 

The Qualcomm Design in India Challenge is in its sixth year. What will be the core tech areas in focus and what kind of challenges would you like to see startups solve?

We're starting our sixth cohort of the Qualcomm Design in India Challenge, and it will culminate early next year. I really want to see as many 5G use cases as possible. I am really hopeful that 5G networks will be in commercial service by this time next year. 5G phones are already in use, with few million units already sold. With networks following devices, non-handset use cases should be ready with the network. My hope is to encourage as many hardware designers in India as possible to innovate with 5G in mind. In the last three years, all our startups have embraced on-device AI. Working with Qualcomm also guarantees strong cellular connectivity. Another major focus area is robotics. For instance, Qualcomm has been working with NASA's Jet Propulsion Labs to power the MARS helicopter. The capabilities of cellular networks can be well extended to augur innovation in space tech and robotics as well. 

The market for innovation in hardware in India is relatively new but looks promising. Can you touch upon the classic challenges and opportunities for companies in this domain? How can India become a hardware innovation hub?

I agree, there is a growing market and appetite for hardware innovation in India. To give you an idea, I would like to highlight some of the successes from the 2020 QDIC cohort - Adiuvo Diagnostics developed a patented multi-spectral, AI-enabled imaging technique for rapid and touch-free diagnosis of a range of diseases; Planys Technologies built underwater drones that function well in low latencies. Our efforts to support these startups were not in isolation - we work with a range of incubators and research organisations like IIT-M, NASSCOM CoE IoT and Kochi Startup Mission. So, startups definitely have an ideal ecosystem to thrive. MeitY too has greatly supported endeavours in hardware innovation. Rolling out multiple policies like the PLI has expanded the scope of local manufacturing. Thanks to this scheme, some manufacturers have progressed from parts assembly to Surface Mount Technology (SMT) - but the question is, can these components be made in India? Can we enhance hardware testing and reliability standards? Can we write software for this hardware and also support it with due security measures? Interconnectedness is how we can truly become aatmanirbhar. I see that the right questions are being asked, and the right people are being approached to solve these critical issues. For us to get close to where China or Taiwan are today, an integrated approach is needed. A good start would be developing dual capabilities in hardware design with software & systems design that will set new standards in Asia, and catapult India as a segment leader.

Qualcomm's strong R&D presence in India is encouraging. What are the biggest strengths of the India R&D centres? What specialised technology and hardware offerings here are poised to take on a future that will be largely dominated by AI?

When we decided to set up an R&D centre in India in the early 00s, we were bolstered by the availability of raw talent. Nearly a third of the 40,000+ strong workforce is based in India, and majority of the company is comprised of engineers. In 2003, we were thinking of design centres and this led us to starting the lab in Hyderabad. While our Bangalore facility focuses more on System on a Chip (SOC) design, the Hyderabad centre emphasises on software and testing. In 2011, we added a team with WiFi expertise in Chennai and in 2015, brought onboard engineers with experience in Bluetooth-based voice in Gurugram. Preponderance of talent has always been the key driver behind building Qualcomm India R&D centres. Other reasons for driving R&D in India were the low incidence rates of IP pilferage and a booming electronics market. 

5G will revolutionise industries in the future, and boost AI's potential even further. How are you planning to leverage this opportunity in India?

I'll be honest here, I do believe India is a little late to 5G, when compared to the overwhelming embrace of 4G in the past. Yet, I do believe this year would be the right time for networks to become 5G-ready and we are anticipating its rollout. While handsets are already 5G-ready, the network itself is built for industries beyond cellular. 5G is designed to provide last mile fibre connectivity. Let me tell you how - if you've seen the PGA tour, cameras have huge cable trays that have to be lugged across a golf course. A 5G-enabled camera solves this problem immediately. Today, almost all EV cars have cellular modems inside. Car manufacturers have inserted the modem in the front, allowing the car to pick up data points as it moves along the assembly line. With 5G, there is practically no need to worry about on-device WiFi linkages, making it ideal for low-resource and dynamic settings. We're impatient for a 5G rollout in India, and we're big proponents of the millimetre wave spectrum because of its wide availability and even wider bandwidth. As India's torrid pace of construction occurs across multiple industries moves, we want to be WiFi-ready, fiber-ready and 5G-ready.