How Meltwater’s Hyderabad Hub is Building AI-Driven Media Intelligence

Media, social and consumer intelligence firm Meltwater recently expanded its AI hub in Hyderabad with a new 14,000 square feet facility. 

The centre currently houses a 60-member R&D team focused on developing next-generation AI capabilities, with plans to scale to 150 engineers by 2026. Earlier this year, Meltwater launched an AI centre in the city with an initial team of around 20 employees.

Meltwater’s Hyderabad hub plays a key role in the company’s AI-driven media intelligence platform, which processes 1.5 billion pieces of content daily across more than 100 languages. 

At the heart of this innovation is Meltwater’s agentic AI framework, MIRA, which deploys specialised sub-agents to perform tasks such as sentiment analysis, entity extraction and media monitoring.

Talking to AIM, Aditya Jami, CTO at Meltwater, highlighted the Hyderabad hub’s contributions, stating, “Some of these complex agent tech use cases…were also built in Hyderabad, contributing to the main product that we have.”

Advanced AI Stack and Model Distillation

Meltwater leverages GPT-5 alongside its own fine-tuned OpenAI models, using proprietary data to optimise performance for specific tasks. The company also relies on Whisper models for speech-to-text processing.

Given the scale of operations, running at a scale of roughly 6,000 documents per minute, the company applies heavy model distillation, which includes converting large general-purpose models into smaller, high-performance, task-specific models. This approach ensures rapid, high-quality insights while supporting time-sensitive operations.

The Hyderabad team has applied this approach to sentiment analysis, entity extraction and name recognition across multiple Asian languages, efficiently adapting large language models for precise tasks.

Independent analysis by Forrester Research shows that Meltwater’s AI can deliver 25-30% productivity gains for marketing, PR and communications teams, reducing the grunt work and freeing professionals to focus on strategy and storytelling.

Highlighting the operational scale, Jami explained, “We have a huge partnership with Microsoft… OpenAI models… We also do heavy distillation, because we have to process 16 billion AI inferences every single day.”

By distilling large models into smaller, specialised ones, Meltwater achieves the speed, cost efficiency and accuracy required to support complex NLP and computer vision workloads.

AI Talent Density in Focus

India has long been seen as a cost-arbitrage destination for global capability centres (GCCs), but the country has decisively outgrown that identity. 

Today, the country’s biggest advantage is not cheaper talent, it’s denser, deeper and more specialised AI talent than most global markets. As a result, GCC strategies are increasingly shaped by where they can find the strongest AI and engineering capabilities, rather than by the lowest costs.

Speaking to AIM, Kapil Joshi, CEO at Quess IT Staffing, mentioned, “High-density AI teams give us faster go-to-market, superior user experiences and significantly higher productivity and scalability. Simply put, the organisations that concentrate AI talent move faster, innovate better and win more.”

This shift is evident in NLB Services’ latest report titled ‘Workforce 2.0 Reset – India’s GCCs Go AI-Native’, which notes that 58% of GCCs have already moved beyond AI pilots. The GCC workforce is projected to reach 3.46 million by 2030, adding 1.3 million new roles, with an 11% jump expected as early as 2026.

Yet, this demand far outstrips supply. TeamLease Digital’s ‘Digital Skills and Salary Primer Report for FY2025-26’ highlights that for every 10 generative AI roles in India, only one qualified engineer is available.

This is exactly the challenge and opportunity that Meltwater navigated while building its GCC in Hyderabad. 

Moreover, describing the rapid pace of scaling, Jami added, “We had nothing when we started from zero… That’s the initial cohort of folks. But that has expanded to 60 over a year.”

Crucially, Meltwater wasn’t trying to recreate traditional engineering teams. Instead, the Hyderabad hub was designed as a high-density AI innovation centre capable of driving global product development. That required hiring specialists in natural language processing (NLP), natural language understanding (NLU), distributed systems, computer vision and large-scale search, rather than building large teams based on cost differentials.

The hub is set to grow to 150 employees by the end of next year, driven by Meltwater’s need for deep expertise in large-scale data systems, NLP, NLU, computer vision and distributed systems.

Speaking to AIM, Rakesh Sancheti, chief growth officer at Tredence, noted that global enterprises are no longer choosing India solely for scale.

“India stands at the epicentre of this shift, not only for its scale but for the depth, cross-disciplinary strength and sustainability of its AI talent pipeline. 

Furthermore, with a rapidly expanding STEM base, strong industry, academia alignment and a culture of continuous learning, India is developing talent equipped to build, operationalise and govern enterprise-grade AI systems with speed and rigour.

Jami also pointed to the strong foundational skills emerging from Indian universities. “In general, India has a really good math background and core engineering background. AI and ML are pretty much like a curriculum now in most undergrad schools at this point,” he said.

Meltwater has leveraged strong university partnerships to supplement experienced talent by bringing in final-year students who can complete their master’s thesis alongside Meltwater.

“You not only hire folks but also get pure R&D projects, which end up converting into some good product ideas,” Jami added.

This blended hiring model, combining lateral talent with university hiring, will continue as the Hyderabad hub scales.

The post How Meltwater’s Hyderabad Hub is Building AI-Driven Media Intelligence appeared first on Analytics India Magazine.