India is emerging as a pivotal player in the AI landscape thanks to its engineering talent and focus on application-driven innovation across sectors. And the biggest theme for next year, without a doubt, is going to be AI agents.
“I’m actually convinced that a faster adoption will be seen in agentic AI. Each one of us will have an AI agent that knows us really well; it will analyse our business and routines and support us in becoming more productive and efficient,” said former CEO of Tech Mahindra and co-founder of AIonOS CP Gurnani at AIM’s MachineCon GCC Summit 2024.
Echoing a similar sentiment, the VP of AI product management at Redis, Manvinder Singh, told AIM that India was perfectly positioned to lead the charge. “The Indian tech ecosystem is going to play a very critical role in agentic AI,” he added optimistically, citing companies like Kore AI and others, which are currently using Redis as a data platform to power their virtual AI agents.
At Redis, Singh leads innovations in vector search, semantic caching, and agent memory. Prior to joining Redis in July 2024, he worked for a decade at Google in AI as director of product management, where he focused on building LLMs, AI frameworks, and other developer products.
He is also a lead SME for Google’s AI Essentials Course on Coursera. “I helped build a partnership between Redis and Google Cloud during my time at Google,” recalled Singh.
A former McKinsey associate partner, Manvinder holds an MBA from Kellogg and a BTech from IIT Delhi. “I grew up in Delhi, and my personal connection to India keeps me deeply invested in its tech ecosystem,” Singh said, emphasising his belief in India’s potential to tackle hard AI challenges and how Redis is perfectly positioned to scale this to a whole new level.
Redis, FTW!
“Redis has always been synonymous with speed like it was the performance database,” said Singh, emphasising its deep-rooted reliability. These capabilities, essential for building next-generation AI agents, address critical pain points like memory, context, and latency. Redis’s integration with Amazon Bedrock, Microsoft Azure, and LangChain also positions it as a developer-friendly choice.
Redis owes much of its performance to its single-threaded architecture—a design choice that has sparked debates but remains a cornerstone of its success. By avoiding locks and minimising system calls, Redis achieves lightning-fast operations. As one user aptly noted, “If you don’t take locks, aren’t making syscalls nonstop, and aren’t fighting cache, you get really good performance.”
While some question the limitations of a single-threaded model, others highlight its simplicity and efficiency as the key reasons behind Redis becoming a favourite for high-speed data workloads. “Redis was built to be single-threaded, which was the right design choice for caching use cases,” explained Singh, “but now we support multi-threading for things like vector search.”
This move has positioned Redis as a formidable competitor to players like Milvus and Qdrant. Unlike these vector-only databases, Redis offers unmatched flexibility, allowing developers to handle multiple data types in a single platform.
“Using a vector-only database is like buying a car that only takes you to the grocery store,” quipped Singh, emphasising Redis’s edge in catering to diverse use cases with a unified solution.
While competitors like SingleStore, Milvus, Qdrant, and hyperscalers offer Redis-compatible solutions, it distinguishes itself with performance and flexibility. It is the most downloaded database on Docker Hub, and its ability to serve as both a caching and vector database gives it a unique edge.
For instance, Asurion, a global insurance provider, optimised API usage by using Redis for semantic caching and routing, achieving a 70% hit rate and significantly reducing costs.
What’s Next for Redis?
In August, the company announced Redis 8, a new update that brings advanced features like JSON, search, and vector databases to its Community Edition. This update is also accompanied by Redis for AI—a package designed to power GenAI applications—marking a major leap in developer access and AI-driven innovation.
Redis told AIM that it is now doubling down on its core strengths—speed, memory, and flexibility—while introducing innovations like Redis Flex, enabling terabyte-scale data handling.
“We’re building new products to solve challenges like semantic caching, memory optimisation, and guardrails for responsible AI,” he shared. Strategic partnerships with AWS Bedrock and Microsoft Azure underline its commitment to becoming a cornerstone of the GenAI ecosystem.The company is also eyeing deeper integration into enterprise AI workloads, with a focus on reducing developers’ friction through tools like LangChain and investments in disk-based data capabilities.
The post 2025 is the Year of AI Agents, and India is Leading the Charge appeared first on Analytics India Magazine.
