Software Leaders Uncensored: Meltwater’s CTO, Aditya Jami, on Building Enterprise-Grade AI Systems

This episode of Software Leaders Uncensored features Aditya Jami, Chief Technology Officer (CTO) at Meltwater, discussing the challenges of building and scaling enterprise-grade AI systems for media intelligence.

Listen to the full podcast here: Software Leaders Uncensored, Aditya Jami on Building Enterprise-Grade AI Systems That Scale

Scale and Strategy

• Massive Data Processing: Meltwater processes an immense 1.5 to 1.6 billion conversations (online news, social, broadcast, and LLMs) daily. This data is enriched using proprietary AI methods (NLP, Computer Vision) and indexed for real-time analysis.
• The GenAI Shift (Mira): Meltwater launched 'Mira,' an AI-powered conversational teammate, to shift the user experience from complex UIs to a simple chat interface, reducing the customer's time-to-value.

Technical Challenges & Solutions

• Relevance: The core technical challenge was moving beyond keyword indexing to vectorizing indices to boost relevance across 100+ languages and various data types.
• LLM Reliability: Jami details the difficulty of making Large Language Models (LLMs) perform deterministic workflows. This required developing strict context engineering and evaluation systems to ensure stability and consistency.
• Infrastructure: The stack relies on Kubernetes and Terraform (IaC) across a multi-cloud environment (AWS and Azure). Data utilizes Elasticsearch, Kafka, and MongoDB.

Leadership Takeaway

The primary challenge for tech leaders is balancing rapid AI innovation with maintaining enterprise-grade stability. Jami advises leaders to design products based on what AI models will be capable of in the near future, not just today.

Meltwater is a global leader in media, social, and consumer intelligence, helping over 27,000 brands worldwide understand and shape their stories across media, social and LLM environments.Learn more about Meltwater.