A daily dose of health and happiness with AI

Can Artificial Intelligence (AI) make people healthy and happy? Abhijit Khasnis, the Chief Technology Officer at Healthify, a Bengaluru-based digital health and wellness platform, believes it can. Among the many goals AI can help achieve, health and happiness stand out. As AI tackles increasingly complex tasks and becomes more capable, Abhijit expects that it will scale up to handle jobs beyond our current comprehension. Health and happiness are among these, with AI tools evolving to become the most reliable and hardworking assistants humans have ever had.

Abhijit explained, “In the health and wellness space, AI will help us build good, healthy habits. We’ll have a virtual assistant in our pockets, guiding us on what to eat and do for our health. Over time, this will reduce hospital visits and improve overall quality of life.”

While Abhijit’s current focus is on health, he envisions a broader impact of AI. “AI’s ability to mimic human behaviour will improve, transforming many sectors, including coding. The way we write, deploy, run, and maintain code will change. As AI enhances robotic capabilities and manufacturing, we will see the rise of ‘universal factories’ — facilities capable of producing anything, much like software can be written to perform any task,” said Abhijit, who studied electronics engineering at Shivaji University in Kolhapur, Western Maharashtra.

The AWS edge

All this will require massive amounts of data crunching, analytics and smart tech running in the background. AWS tools fit the bill. “We have been using AWS for over a decade now. We use SageMaker to build our own custom models. These are not large language models (LLMs), but proprietary models so that our insights are better than what a typical LLM can give. We also use AWS Bedrock”, he shared.

AWS SageMaker is a cloud-based machine-learning platform that allows the creation, training and deployment of Machine Learning (ML) models on the cloud. It can be used to deploy ML models on embedded systems and edge-devices. While Amazon Bedrock is a fully managed service that offers a choice of high-performing foundation models (FMs) from leading AI companies like AI21 Labs, Anthropic, Cohere, Meta, Mistral AI, Stability AI, and Amazon through a single API, along with a broad set of capabilities that can be used to build generative AI applications. 

With Amazon Bedrock, users can experiment with and evaluate top FMs for their use case, customise them with their data using techniques such as fine-tuning and Retrieval Augmented Generation (RAG), and build agents that execute tasks using their enterprise systems and data sources. Being serverless, Bedrock eliminates the need for infrastructure management, allowing users to seamlessly integrate and deploy generative AI capabilities into their applications using AWS services.

Ria: a virtual assistant with real solutions

Abhijit has been a stalwart in the technology sector for 25 years. Before joining Healthify three years ago, he held key roles at Yahoo, Oracle, and the Tata Group. At Tata, he was instrumental in launching their healthcare initiative, Tata Health.

In the mid-1990s, while studying electronics engineering, Abhijit’s interest in computing was sparked by coding projects aimed at optimising public transport routes. His career took off at Yahoo, one of the first internet giants with millions of users, where he honed his skills in data management and coding. Abhijit was on the media side, ensuring personalised, unduplicated news feeds for Yahoo users. Later at Tata Health, he became involved with building ML based systems that mapped symptoms to disease through the right diagnosis. This innovation ensured doctors had comprehensive patient health backgrounds before consultations.

Currently at Healthify, Abhijit is deeply involved with Ria- the AI wellness assistant. Ria provides personalised coaching on nutrition, fitness and overall well-being. Ria can offer workout and dietary recommendations as well. “There is a lot of data collected to help Ria make the decision to correlate a person’s current health with food and workouts. The fun thing about Ria is, it can also see,” said Abhijit. A user can take a picture and share it with Ria and the assistant can also talk in natural language, similar to normal conversations between human beings.

Ria handles a considerable amount of back-end processing and offers proactive nudges on the right wellness program for users. “What was science fiction earlier is becoming reality in many cases. AI has evolved from doing repetitive tasks, to doing creative tasks and now strategizing and decision making as well. So, AI has expanded in scope and capability over the years,” he reflected.

AI: an inevitability

Over time, AI will become the default, much like people use the internet and smartphones today. Users won’t have to mention that they are on AI to do tasks, it will be in the background, helping in multiple ways.“Every conversation that I have with the engineering, product or design teams revolves around all the new things that are happening. AWS also provides platforms to keep up-to-date with new developments in the field, said Abhijit.”

Abhijit sees some of the old problems, like hunger, being solved and new ones, like obesity, coming up. These will be tackled with the use of AI. “Major shifts have happened — from getting enough food to getting the right kind of food. Health consciousness is increasing. We see people engaging with virtual agents like Ria to address problems such as being overweight”, he shared.

Apart from Ria, Healthify also offers Snap. Users upload thousands of food images on Healthify Snap to get real time information on the nutritional benefits of the food they are consuming. Healthify Snap was originally built with AWS infrastructure before expanding in its capabilities with vision-large-language-models from leading providers. The proprietary models built on AWS complete the ensemble of processing systems that analyze the images. It can handle tens of thousands of images every day. “We have had faith in AWS for more than a decade. It offers rock solid cloud infrastructure, reliability, availability, and performance has never been an issue,” said Abhijit.

As users grow comfortable with AI assistants and as technological capabilities enable more tasks to be automated, Abhijit sees AI becoming more intuitive in the coming years. “ It will become natural to interact with AI. We will shift to more conversation-based AI interactions as compared to command-based at present,” he said. AI systems will understand contextualities and emotions and respond accordingly. It will comprehend changes in voice, mood, expression and adjust its response accordingly. Reflecting on these changes, Abhijit said, “It will become natural for people to interact with AI. I think it will be a lot more human-like,”

As for the impact of AI on society, Abhijit sees the emergence of an `AI first generation’ that will expect virtual assistants to augment their daily lives. The movement that began with wearables, monitors, smart watches, fitness trackers, is now enabling users to understand more about themselves and what they do. AI will gradually continue to make people more aware about their surroundings, inherent risks and the opportunities available. The future will be AI-led and smart.