Singapore's Mass Rapid Transit (MRT) system is one of the busiest in the world, moving millions of commuters daily across a network that spans hundreds of kilometres. Keeping the trains running safely and on time requires a herculean effort, especially with only a three-hour window each night for maintenance crews to access the tracks and fix faults. Now, the country's second-largest rail operator, SMRT, has deployed a new weapon in this ongoing battle: an artificial intelligence platform called Jarvis.

Unveiled at the Oracle AI World Tour Singapore on 14 April 2026, Jarvis — which stands for “Just Another Really Intelligent System” — was co-developed by SMRT’s engineering arm Strides Technologies and Oracle Corporation. The platform brings together decades of operational data in a unified analytics layer, using Oracle Cloud Infrastructure (OCI) Enterprise AI and the Oracle Autonomous AI Database. The result is a system that not only speeds up fault identification but also enables predictive maintenance, potentially preventing failures before they occur.

How Jarvis Works

The core of Jarvis is the consolidation of more than 30 years of SMRT’s historical data, previously scattered across multiple systems in various formats such as text logs, graphs, and flowcharts. This wealth of information — covering engineering records, failure patterns, maintenance logs, and operational timelines — is now accessible through a generative AI chatbot interface. Powered by large language models (LLMs) and vector search, the chatbot allows maintenance engineers to ask natural language questions and get precise, actionable answers in seconds.

For instance, an engineer might query about recurring faults in a particular type of point machine — the mechanical device that controls track switching. Jarvis can instantly translate the textual and graphical records into exact geolocation coordinates, indicating which specific point machine along the permanent way requires attention. This eliminates the need for technicians to manually search hundreds of kilometres of track, saving precious hours during the nightly maintenance window.

“Suppose you are aware of certain faults that have been occurring. Now you need to translate that to exactly which point machine on the permanent way is acting up,” said Ngien Hoon Ping, Group CEO of SMRT. “They go directly to the point machine that same night window and deal with it. It achieves better effectiveness, high productivity and cost-savings.”

Predictive Maintenance and Reliability Metrics

Beyond pinpointing existing faults, Jarvis uses machine learning algorithms to analyse failure patterns and predict when equipment is likely to fail. This predictive capability is critical for improving SMRT’s Mean Kilometres Between Failure (MKBF), a key performance metric that measures the average distance a train travels before experiencing a service-affecting fault. The Land Transport Authority (LTA) mandates a strict MKBF target of one million train-kilometres, and any shortfall can lead to service disruptions that inconvenience millions of commuters.

By proactively identifying components that are approaching the end of their operational life or showing early signs of wear, SMRT can schedule replacements during routine night shifts rather than reacting to emergency breakdowns. This shift from reactive to predictive maintenance reduces downtime, extends asset life, and improves overall network reliability. It also optimises the use of maintenance crews, who can focus on high-priority tasks identified by the AI instead of following a rigid schedule.

Oracle’s Role and Technology Stack

Jarvis runs entirely on Oracle Cloud Infrastructure, leveraging the Autonomous AI Database to handle the heavy lifting of data ingestion, indexing, and querying at scale. Oracle’s Enterprise AI suite provides the building blocks for the LLM-powered chatbot and the machine learning models for fault prediction. According to Chin Ying Loong, Oracle’s senior vice-president and regional managing director for ASEAN and South Asia growing economies, the partnership demonstrates how enterprises can bring AI to where their data resides.

“Rail operators depend on timely, accurate data to keep services running safely, reliably and on schedule for millions of commuters each day. Running on OCI, Jarvis demonstrates how Oracle can help bring AI to where enterprise data resides to improve efficiency and operational responsiveness,” Chin said during the launch event.

Impact on the Workforce

Despite the high level of automation, SMRT’s leadership is emphatic that Jarvis is designed to augment human workers, not replace them. The company continues to hire engineers, electricians, and technicians even as it expands AI capabilities. The platform is currently in its first phase of deployment, with over 50 SMRT engineers actively participating — some are cleaning and tagging data for AI training, while others are coding AI agents that automate routine analysis tasks.

“SMRT is still hiring, even in the face of this AI world. We still need engineers,” Ngien stressed. “To us, AI is really about enabling the organisation to uplift our people.”

This philosophy aligns with Singapore’s broader national strategy of adopting AI to enhance productivity while ensuring that workers are reskilled for higher-value roles. SMRT’s internal “Kaizen” culture of continuous improvement — a Japanese concept popularised by Toyota — has been key to integrating AI tools into daily workflows.

Future Plans and Global Sharing

Looking ahead, SMRT plans to extend Jarvis to cover more asset types beyond point machines, including signalling systems, power supply units, and rolling stock components. The company also hopes to share its AI models and best practices with other rail operators globally. Many metropolitan rail systems face similar challenges of aging infrastructure, limited maintenance windows, and rising passenger expectations.

“They also have a trove of data, so through the models we’ve developed with Oracle, we would be happy to share with other operators,” Ngien said. This open approach could accelerate the adoption of predictive maintenance in the railway industry, potentially improving safety and reliability worldwide.

In a world where urban populations are growing and public transport networks are under increasing pressure, innovations like Jarvis represent a tangible step toward smarter, more resilient infrastructure. By combining decades of domain expertise with cutting-edge AI, SMRT is not just keeping trains on time — it is redefining what’s possible in rail maintenance.

Source: ComputerWeekly.com News