How Talent, Technology, & Technique Help Build Future Businesses
I believe that we have to build technology as an integral part of delivery. In a recent interaction with the Editor of siliconindia, Radha Basu, Founder & CEO, iMerit, shared her insights on how the role of People, Processes and Technology are high on the list of topics.
Radha Basu has a Master’s in Computer Science and Biomedical Engineering from University of Southern California and has done an Executive Management Program in Business Administration and Management, General from Stanford University Graduate School of Business.
Radha founded Santa Clara University’s Frugal Innovation Hub and co-founded the Anudip Foundation - iMerit's sister foundation. She also serves on the boards of NetHope, Jhumki Basu Foundation and the Miller Center for Social Entrepreneurship.
Radha has received many honors including the Global Thinkers Forum Award, UN Women-ITU Gender-Equality Mainstreaming Technology Award, Silicon Valley Business Journal Women of Influence Award, India's Best Leaders in Times of Crisis recognition by Great Place to Work, Top 25 Women of the Web and CEO of the Year.
1. Given your experience, what is the role of People, Processes and Technology and how have these evolved in terms of becoming an integral part of organization across India?
People, processes and technology are high on the list of topics, especially if you look at artificial intelligence (AI) and Machine Learning (ML). These are entering the operations side as well as the data and model side. I look at the words slightly differently, and the meaning has evolved too. We look at it as technology, talent, and technique. As my experience goes back to building enterprise solutions, I had a very large team with around $1.2 billion in business enterprise solutions. We had technologies, which were being developed in different kinds by ourselves. Accordingly, we would look at what kinds of technologies were needed to work on it and then train the people. Then we would look at people and the process to understand what should be the workflows and the kinds of people working on it. We looked at a problem and evolved the workflows and processes accordingly.
After putting the solution together, you would monitor that and find some iterative loop. But essentially, other than improvements in the technology and getting people even more experienced at it or maybe working the process a little bit, you stayed with it, implemented, and deployed it.
Once you work on a particular business or technology process for a while, you iterate through it to improve it further. A lot has changed quite dramatically in the last six months. I have observed that we're starting to look at AI and ML as an integral part of enterprise digital transformation. Previously, there was code, and you would run the data through it. We use the data to develop the models and run the data through it in deployment. It impacts the models a lot. Hence, the process or the technique is a very tight integration of data collection, ingestion, labeling, mapping, quality, auditing, and then running it back into the model. Therefore, when you look at any AI deployment or ML, ML Data, or ML DataOps, you generally have the model-centric and data-centric approach.
It is because AI is starting to get implemented, and we are spending a lot of our time here. For me, the changes from running large enterprise solutions organizations to what we're running and scaling today are two very different things. The iterative nature of the process requires a fusion of technology - the ML and AI technology, the models, and the intelligence of people.
It is almost like an infinite loop. I have never witnessed such a tight infinite loop in the 45 years of being in the industry and scaling really large organizations.
I believe that we have to build technology as an integral part of delivery. There is technology in the products and the models, in operation and the deployment, as well as in the intelligence of the people. All the three come together – technology, talent, and technique- as a fused loop.
2. How important is it for an organisation to realise the potential of bringing these three together, especially the potential it brings in India?
I think that people are starting to realize it. We had 3000+ registrants on the first ML DataOps Summit. Many people, and several participants from India, were from the enterprise organizations, implementing digital transformation, AI, like Chief Digital Officers, Chief Information Officers, and CDOs. Hence, they are interested, and there are also tech companies who have developed models. We are going to have such a profusion of data and this then ties back into being able to validate, audit, and monitor the data. We need to develop new policies, a technique around training the models, and algorithms that will change.
As in my keynote, I would say, "it's the first baby steps towards the entire area of getting AI into production."
For instance, at an recent event, people from TCS, Mindtree and several other organisations, talked about how they are putting together almost a separate company, a unit to focus on AI production. It's because it demands different thoughts and collaboration to develop this. I have seen the fusion of the three things to get digital transformation moving faster. In India it allows us to create many new jobs due to the requirement of the new skills, new technologies, keeping on top of the technologies with new techniques, process people, and L&D people
For the beginner, I have always believed that you incubate things and innovate in small ways over the years. In these pods, you can incubate and innovate with a group of people who understand problems like solution architects, and solution consultants. The process of incubating and innovating is to be completely ready to disrupt yourself. It is because if we do not disrupt ourselves, someone else will disrupt us.
3. So keeping in mind your ‘technology, talent and technique’, what is the impending evolution in the digital transformation journey?
Process is part of the techniqu, because the technique has processes and policies that trains the algorithms. Policies, processes, and skilling are all part of the technique. It is the technique of how the technology and the talent make these products work. So, when I say that it has evolved to these three, it is a very early stage of seeing these fuse together. And I think we have a lot of work to do to make this all work in a scalable business model with infrastructure.
Applications of digital transformation, AI and ML, particularly, are just starting to come out in various segments. Therefore, the next step would be taking these three pieces and taking AI into many different applications. Additionally, scalability and security will be two of the very key things that we have to focus on. It's because in all of these digital applications in the transformation, particularly of AI and ML, security is going to be extremely critical due to the availability of so much data.
In our summit session, ‘2022:'The Year of ML DataOps' – The Ground Truth of AI’, we represented the idea of taking AI into production. The summit was all about presenting why machine learning data operations play an integral role in bringing artificial intelligence to market at scale and discloses why 2022 is shaping up to be the 'Year of ML DataOps.' A concluding fact of the summit is a feedback loop of results will constantly force enterprises to adapt their ML data operations to meet the demands of their models. Algorithms in the field will come back with edge cases, which data operations will scramble to solve, before the algorithm is redeployed.
Factors that assist increase annotation efficiency consist of expertise on nuances and edge cases, work experience, motivation for the job, cross-trained annotators, confidence to challenge algorithms and provide insights, diversity of opinions and backgrounds, and accountability and transparency.
Combining technology and human-in-the-loop expertise gives enterprises a true end-to-end solution as they move to deploy their models in the field. The highest quality possible data will be generated by bringing together the right expertise, judgment, and technology.
So, I feel we are in the nascent stage, and we have to perfect that to scale in a secure manner. And I think it will take us some time to do it. So that's what I want to focus on.
The digital transformation witnessed in the past 5-10 years is around the code, people, process, and technology. Now, we have an infinite loop of data in the middle of the models and the algorithms that will go into the products and drive the transformation. On a concluding note, I would say it is high time to look at the highest quality data that drives the most successful digital transformations, particularly using AI & ML.