India boasts a staggering Redis user base of approximately 12 million downloads per day, securing its position as the third-largest adopter on the planet, closely trailing behind the technology giants, the United States and China.

Amid this expansive user base, Bengaluru (rightly referred to as the Silicon Valley of India) emerges as a distinguished hub for Redis adoption with the highest adoption rate globally, followed by Mumbai and more.

The recent surge in Redis adoption across India has been the result of the boom of generative AI.  The country actively engages in advancing its capabilities in this domain, recognising the pivotal role that the database plays in shaping these foundational models.

During Redis’ cofounder and chief tech officer Yiftach Shoolman’s recent visit to India, AIM caught up with the visionary tech veteran to understand the application of Redis by Indian companies, shedding light on the future plans of the Redis team, addressing the prevailing AI hype, and more.

How Indian Firms are Using Redis

“Every month, half a billion people use Redis Enterprise in India alone,” Shoolman told AIM. 

Adding on to what he said, Dishank Nagpal, country manager head of Redis further explained how a diverse set of companies, ranging from media to trading are leveraging Redis with some case studies. Groww, Purplle, Apna, AngelOne, and Zee Entertainment are some of its customers. 

For example, job search platform Apna which caters to tier-two and tier-three markets manages the creation and storage of job seekers’ profiles and provides real-time feeds of relevant job opportunities based on factors like location, experience, and availability with Redis. This real-time approach has proven effective in connecting blue-collar workers with suitable employment, addressing the challenge of finding relevant jobs in these markets.

Many blue-collar workers lack internet experience but possess specific skills. So using Redis, Apna has developed an AI model that automatically generates sentences describing the abilities of these workers, aiding in articulating the capabilities of first-time internet users in a fully automated manner, leveraging Redis for search, personalisation, scalability, and augmentation.

Additionally, another customer, Zee Entertainment, a media house, faced significant challenges operating across five regions. They employed Redis Enterprise’s active-active solution for distributed caching, allowing updates made in one region to be automatically reflected in the other five. This not only streamlined operational efficiency but also resulted in a substantial cost saving of around 70%. Redis Enterprise’s capabilities proved essential for Zee in overcoming manual tasks and achieving a seamless operation across diverse regions.

Redis is also working with a Mumbai-based full-service retail brokerage firm. The company’s system encountered delays while accessing their investment portfolio, around 480 milliseconds to a second for each share, disrupting the seamless experience of exploring and making decisions on shares. 

By transitioning from open source to Redis Enterprise, they experienced a remarkable improvement, reducing latency from 480 milliseconds to just 20 milliseconds, even with a billion operations per second. This transformation showcases the tangible benefits and performance enhancements achieved by adopting Redis Enterprise for critical applications.

“This issue showed how important real-time solutions are, emphasising the contrast between a sluggish, reminiscent-of-the-80s experience and the efficient, responsive user experiences facilitated by Redis,” commented Shoolman. 

Backbone of Generative AI

Databases form the backbone of large language models. “We have been working with vector databases even before generative AI came into action,” Shoolman noted.

Redis is not only providing real-time data to fuel the generative AI wave but has also collaborated with LangChain to release OpenGPT, an open-source model that offers a flexible approach to generative AI, allowing users to select models, control data retrieval, and manage data storage. 

“If you observe, GPT has a somewhat limited window function. While it’s fantastic to work with, the accuracy of its output may be less than optimal, depending on the task,” said Shoolman. 

And this is what OpenGPT is trying to solve. It allows for the selection of different models, extending beyond the confines of GPT. Furthermore, OpenGPT facilitates interaction with data across multiple domains. 

“LangChain is known for its flexibility but requires a higher skill level from developers. It goes beyond simple file uploads, allowing developers to decide on the structure of knowledge bases, whether they are based on general information or user prompts,” he added. Developers can create their knowledge bases of prompts and add relevant content to requests associated with these prompts. 

What’s Next for Redis

Shoolman believes the future appears positive and promising for the company.

“Examining trends impacting real-time aspects such as network agility, device performance evaluation, three-dimensional interfaces, immersive experiences, and the pervasive influence of AI, it’s evident that the factors driving the need for a real-time database are converging rapidly,” he concluded.

Read more: How Redis is Fueling the Generative AI Wave