Over the last five years or so, there has been a rising trend of global tech heavyweights investing heavily in India’s booming AI ecosystem in the form of accelerators, research centres, and academic partnerships; a testament to the country's flourishing AI talents as well as growth opportunities. The Bosch Centre for Artificial Intelligence in Bengaluru is one such notable example.

Established in 2017 as the company’s third AI research centre, after Sunnyvale in the US and Renningen in Germany, the centre has been at the forefront of India’s AI research space. The company has also announced its plan to invest over €300 million in the centre in the coming years.

Nataraju Vusirikala, a seasoned data scientist who has crafted his skill in the US and Germany, heads the centre in Bengaluru. With decades of experience in AI consulting, product development and project delivery experience for a wide range of industry verticals like automotive, supply chain and logistics, manufacturing, agriculture and IoT, Nataraju’s current assignment is to identify potential business opportunities for applying AI across different Bosch products and services. 

“The main vision and goal of the Bosch Centre for AI is to work with different Bosch product divisions. It could be a component in the car, it could be a sensor that goes into your phone, it could be a washing machine, or it could even be a heating system,” Nataraju said to INDIAai.

“What we do in the AI division is, firstly we try to see how AI can be a part of the product so that we can enhance the product with features developed through AI. Secondly, we look at how AI can make the process of making the product more efficient,” he explained. “Going forward, it will be surprising to see if there will be even one product where AI will not play a role either in the making of it or as the part of the product itself.”

One of the fundamental objectives of the Bosch Centre for AI is to work with multiple divisions of the company, with a focus on research to apply and to deploy AI-based solutions. 

“AI cannot work as a standalone entity. A certain amount of data and domain expertise which comes from different Bosch product divisions is combined with data mining, and machine learning expertise within our center. Based on that, we develop the solutions on top of the existing and future products. Our competency is largely around algorithms, managing large scale data, deploying AI based solutions and taking care of certain aspects corresponding to AI-based features,” noted Nataraju.

When it comes to AI projects, the Bosch Centre for AI at Bengaluru has a two-pronged approach. The first is to focus on industry verticals or domains, and the second is the development of AI systems that will become an enabler for progress in those domains. 

In terms of “domain, we look at automotive, manufacturing, engineering, IoT, and we also look at the whole supply chain management and as well as something like sales. There are many AI-based solutions that can help all these domains,” he added.

“The other dimension is focusing on the AI aspect. It can be classic machine learning, or it could be more like deep learning, either in terms of computer vision or natural language processing. We also look into unsupervised learning as well as reinforcement learning. Then we do a lot in terms of multi-objective optimization. We take all of these, and the interplay between all of them is what brings AI to life,” added Nataraju.

One of the best examples for this approach is a solution for greenhouse farming called Plantect, where the centredeveloped AI system that analyse data points such as temperature, humidity and solar radiation to provide the right advice to the farmers, for disease prediction and prevention. 

“How we use this data to bring some benefit to the farmer, such as instructions to take necessary correction, is the important aspect. In Plantect, we use IoT alongside a combination of deep learning and classic machine learning to help the farmer in early decision making to prevent loss of the yield,” explained Nataraju.

As of today, Plantect has been deployed in several greenhouses in Japan. The company also have plans to brining similar technology to other countries like India.

“Whatever solutions we develop, we try to customise it for the respective local market. Even though we cannot reveal the actual date and time; there is definitely a plan to expand it to other Asian countries. In addition to this, we are looking at domains like connected services and other areas where we can bring AI to add value to the existing Bosch products,”

Another exciting initiative from the Bosch has been the research partnership established with some of India’s top academic institutions such as IIT Madras. “We have an association with IIIT Hyderabad as well, where we work on other new-age technologies and also have continuous exchange via internships. We have very recently signed with Nirma University, and we are constantly looking at how we can support the community.”

In terms of the COVID19 fight, Bosch was in the forefront, as many prominent Indian AI companies. “We have a healthcare product called Vivalytic which leverages AI. In this case, we collect certain specimens from the people; it could be phlegm or blood. The specimen is then placed on a cartridge with precomputed chemicals, and these chemicals react to the specimen. And based on this output we can do the diagnosis, if the infection is positive or negative.”

When it comes to the question of where Indian AI ecosystem stands in comparison to other countries in the world, Nataraju is realistic yet optimistic. 

“I do not think we can compare with the US or China or others, but in some areas, we are quite advanced. Talk about e-commerce or digital marketing space, or banking; we are quite good in these verticals.” 

“But when we talk about Industry 4.0, IoT, autonomous driving, I think the ecosystem is still nascent. There is definitely a scope to leverage AI for social benefits in India,” he noted

“In general, if you get started with AI deployment, there will be multiple phases such as problem identification, formulation, implementation and deployment. Amongst these phases, we have an ecosystem of talent in India for the implementation phase and the deployment phase. But with the phases of formulating a problem or even identifying the problem, I think our education system and our ecosystem are lacking behind.”

“In India, the focus should be on problem formulation and identification. It is imperative, as the solution that we provide is only good in certain spaces. If we extrapolate, it will not work in some spaces. We need to identify the problems to focus on.”

With new developments happening at a rapid pace, unprecedented in human history, AI research has its own challenges.

“The developments are taking place at a faster pace. Be it image recognition, be it understanding the natural language, be it around automated driving technologies. In the parallel, the data sets are also becoming open. So we continuously monitor these developments within our teams,” Nataraju said.

According to Nataraju, maintaining good communication is the most important challenge of his role. “In the end, when you develop the solution and have it deployed, you need to have a very good understanding of the users of these technologies. And for us, currently, these users predominantly are out of India. So the challenge is, how you maintain good communication with them.”

“Since there is a domain aspect as well as an AI aspect, we should be able to connect to the customer, or the final user to understand what is the status quo process and how our solutions will be integrated to their process and benefit them. The challenge is to bring both these worlds together. And for that we use various tools to make it more transparent and trustworthy between the customer and our team”, concluded Nataraju.