Smita's Tech Odyssey with AI

When Smita Ojha was a student at IIT Bombay, pursuing BTech in computer science, she had her first major interaction with computers. “That time technologies like recommendation engines were just starting to take shape,” recalled Smita, now Senior Vice President of Engineering at Mindtickle, a revenue enablement company. That period also saw Amazon emerge on the global scene and put some of these technologies to improve real world experiences.

Her four years at college, from 1996 to 2000, not only solidified her computer science fundamentals but also instilled a fearless curiosity to explore new technologies. Driven by a desire to delve deeper into new ideas like collaborative filtering, Smita pursued research at the Georgia Institute of Technology in Atlanta. Collaborative filtering, an information retrieval method, recommends items to users based on the preferences and behaviors of similar users.

“Innovations like collaborative filtering, recommendation engines, and data warehousing were still very new. What fascinated me about Amazon was their advancement of collaborative filtering algorithms. That was also my first exposure to Artificial Intelligence (AI) and Machine Learning (ML),” Smita shared.

Searching for new innovations and tech disruptions

Smita quickly moved from theoretical innovation to applying tools like ML in real world use cases through her work in the industry. She applied her research in enterprise search ranking algorithms and real-time streaming/data warehousing solutions. A lot of her early research involved crawling the web in a controlled fashion - not for a generic search query, but to achieve a specific goal. Her early interest and work in these areas led to stints at enterprise software and social media companies.

Smita also embraced difficult engineering problems like moving office 365 from single tenant CD solution to multi-tenant cloud solution and building the world’s largest social graph cache infrastructure serving multi-billion queries per second.

Her passion for technology, innovation and bringing impactful tech-led change to organizations made her successful in multiple top tech companies in the world.

Learnings from those early tasks laid the foundation in the evolving spaces of enterprise search, collaborative filtering, and more. Later in 2010, at a social media firm, Smita led the social graph caching core infrastructure team, responsible for serving seven billion queries per second. This again exposed her to cutting-edge technologies and provided the knowhow of core computing infrastructure.

After nearly 18 years in the US at some of the largest software and social media companies, Smita moved to India to catch the wave of technological innovation happening here. She joined the startup (now a Series `E’ unicorn) Mindtickle, which specializes in sales enablement. It is an all-in-one platform for training, content, digital sales rooms, AI role plays, and more. Mindtickle empowers 300+ organizations, including Cisco, NetApp, Ford, and Sherwin Williams, to maximize the productivity of their sales teams. Sales enablement services aim to maximize revenue growth across the entire customer journey, by equipping customer facing teams, including sales and marketing, with necessary tools, training, content and data to effectively engage with prospects.

The wave of innovation has different adoption patterns across enterprise. While large enterprises were not the first adopter of cloud infrastructure, they are at the forefront of adopting GenAI innovation to improve the efficiency and productivity of their organizations and provide value to their customers. For instance, “Amazon Web Services (AWS) saw the market shifts and was an early mover in the cloud space, opening up a whole new set of opportunities, with AI/ML applied to the cloud,” said Smita, now based in Bengaluru.

Letting a Thousand `AI’ Flowers Bloom

The shift to cloud freed up enterprises from the constraints and limitations of their own data centers. As has been well documented, cloud computing helped companies reduce costs, use `pay-as-you-go’ models, increase scalability to meet fluctuating demands, improve flexibility in resource allocation, enable faster time to market, and so on. Those benefits are enabling quicker deployment of AI models and new applications.

Discussing the major shifts to the cloud, Smita said, “AWS is a great company that shaped the cloud era. AWS laid the foundation for everyone to get onto the cloud. It's not just about hosting; it's about having the right level of backups, resiliency, log structures, monitoring, alerting, and so on.” The tech behemoth’s early focus on solving user problems and meeting all their requirements has indeed been helpful across enterprises, large, medium or small.

In her 22-year career, this visionary leader has seen many technologies come and go, but she views AI as fundamental and here to stay, changing things for the better.

“In the revenue enablement space, we've seen a big change. It used to take three to four months to onboard a new salesperson. Now, we do it in one to two weeks! This is a real change. The salesperson no longer needs to take 10 courses to be ready for the first sales pitch. Instead, they just need an interactive roleplay with an AI bot to prepare them. That's a drastic change in the industry. This doesn't mean we've tackled everything, but we're in a phase of AI where we can let a thousand flowers bloom, and see what impact it can have.”

This change is evident at Mindtickle, where sales organizations can be productive throughout the entire cycle, from onboarding to making a sales pitch to closing deals. “It's about applying the right tools. For instance, in meetings today, nobody expects to take notes or action items manually. That's automatically done, making my job easier. Similarly, I expect anomaly detection algorithms to kick in and inform me if something is going wrong. Today, AI is enabling a lot of things,” explained Smita.

Unlocking impact with AI

While that impact has been felt across multiple domains, Smita only sees it improving from here on - be it in search, coding, video making, picture editing, writing emails, research and so on. Interestingly, AWS has been a partner in this journey throughout. For example, Smita pointed out, “On AWS Bedrock, we highly depend on Claude Sonnet models, which help us carry out self-guided programs and sales training. Our conversational AI (Call AI) uses AWS transcribe and Amazon Rekognition services”.

Amazon Bedrock is a fully managed service that offers developers access to a wide range of foundation models, from leading AI companies, along with the tools needed to build and deploy Gen AI applications. The integration of Claude 3.5 Sonnet into Amazon Bedrock combines Anthropic's (an AI research & development firm) cutting-edge AI technology with AWS's cloud infrastructure and development tools.

Mindtickle also uses AWS Rekognition, a cloud-based computer vision platform. This makes it easy to add image and video analysis to applications. Users can perform image analysis to detect objects, concepts, people, scenes, and also detect inappropriate content.

At Mindtickle, Smita and her team are enthusiastic about these technologies and are among early adopters of any new product or service. Constant learning and exposure to the capabilities of new technologies makes it easier to use and deploy them. “AWS is bringing us the best of breed solutions. We work with their engineering teams as well. They are very customer focused. What I really enjoy is our team taking risks, going ahead with proof of concepts. We have a strong culture of using demo sessions as a learning opportunity,” said Smita.

Another of Smita’s passions is sustainability. She remains conscious of the environmental impact of using energy-intensive AI. Recalling a dinner conversation with Werner Vogels, CTO of Amazon, she shared, “he had similar concerns around the carbon footprint of each query. The good thing is that Amazon is addressing such issues and these are top most on his mind.”

At the same time, we also have to govern the use cases of AI to ensure advancement of mankind, she noted. Looking ahead Smita would like AI to be used carefully to make us productive and efficient across sectors - from farming to pharmaceuticals, edtech to IT productivity.