Keeping trains humming along safely and smoothly across Singapore's rail network is a monumental task, especially when engineers have only a three-hour window each night to fix track faults. Now, rail operator SMRT has a new artificial intelligence (AI)-powered tool to help: Jarvis.
Playfully dubbed 'Just Another Really Intelligent System' by SMRT staff, the intelligent analytics platform was developed by Strides Technologies – SMRT's engineering and tech innovation arm – together with tech giant Oracle. Announced at the Oracle AI World Tour Singapore on 14 April 2026, the platform leverages Oracle Cloud Infrastructure (OCI) Enterprise AI and the Oracle Autonomous AI Database to consolidate more than 30 years of operational, engineering and failure pattern data.
This vast trove of data, previously distributed across multiple systems in the form of text, graphs and flowcharts, is now accessible to maintenance teams through a generative AI (GenAI) chatbot interface powered by large language models (LLMs) and vector search to help them make better-informed decisions. The result is a system capable of supporting predictive maintenance using machine learning algorithms, enabling faster fault resolution and contributing to SMRT's mean kilometres between failure (MKBF) metric, an industry-standard used to measure rail service reliability. In Singapore, the Land Transport Authority sets a strict MKBF target of one million train-kilometres, a benchmark that public transport operators must consistently meet to ensure minimal commuter disruption.
During a discussion with Chin Ying Loong, Oracle's senior vice-president and regional managing director for ASEAN and South Asia growing economies, SMRT group CEO Ngien Hoon Ping said one of Jarvis's biggest benefits is its ability to convert textual and graphical information into precise geolocation data. '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,' he said, referring to the mechanical devices used to control and switch railway tracks. Instead of technicians searching across hundreds of kilometres of track to find the faulty equipment, Jarvis allows them to pinpoint the exact location. 'They go directly to the point machine that same night window and deal with it,' Ngien said. 'It achieves better effectiveness, high productivity and cost-savings.'
Despite the growing use of AI, Ngien stressed that the technology is meant to improve the effectiveness of SMRT's workforce of more than 10,000 people, not replace them. 'SMRT is still hiring, even in the face of this AI world. We still need engineers,' he said. 'To us, AI is really about enabling the organisation to uplift our people.' Jarvis is currently in its first phase of deployment, with more than 50 SMRT engineers actively participating in the process. Some are analysing existing data, while others are involved in coding AI agents.
Ngien noted that managing a complex locomotive network, from signalling and power systems to railway tracks, requires a Kaizen culture of continuous improvement. 'It's a very challenging task, even for the engineers among us. But we have this culture to keep improving and make use of the tools available,' he said. Chin added: '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.'
Moving forward, Ngien said SMRT hopes to share its experience with other rail operators facing similar challenges. '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,' he said.
How Jarvis integrates with existing rail operations
The Jarvis platform represents a significant shift from reactive maintenance to predictive maintenance in rail operations. Traditionally, rail maintenance teams rely on scheduled inspections and historical failure data, which can be inefficient and time-consuming. With Jarvis, machine learning models analyse real-time sensor data from tracks, signalling equipment, and trains to identify patterns that precede failures. This allows engineers to intervene before a fault occurs, reducing unplanned downtime and improving overall reliability.
One of the key challenges in rail maintenance is the sheer volume of data spread across different systems. SMRT has accumulated over three decades of operational records, including maintenance logs, incident reports, and engineering drawings. These were stored in various formats, making it difficult for engineers to quickly access relevant information. Jarvis uses Oracle's autonomous database and vector search to index and retrieve this data through a conversational AI interface. Engineers can ask questions in natural language, such as 'What were the common faults on the North-South Line last month?' or 'Show me the maintenance history of point machine X.'
The platform also integrates with geographic information systems (GIS) to map fault locations precisely. This is particularly valuable in a dense urban network like Singapore's, where tracks run through tunnels, viaducts, and underground stations. By combining text, images, and geospatial data, Jarvis reduces the time needed to identify faulty equipment from hours to minutes.
Oracle's role and technical underpinnings
Oracle provided the underlying cloud infrastructure and AI tools to build Jarvis. The platform runs on OCI, which offers low-latency access to GPU clusters for training and inference. The Oracle Autonomous AI Database automatically manages data security, scaling, and performance, allowing SMRT's data science team to focus on model development rather than infrastructure maintenance. The use of large language models enables the chatbot to understand complex queries and provide context-aware answers.
Oracle's senior vice president Chin Ying Loong highlighted the importance of data location: 'Bringing AI to where enterprise data resides is crucial for responsiveness and compliance. With OCI's distributed cloud capabilities, SMRT keeps its sensitive operational data within Singapore while leveraging global AI innovations.' This is particularly relevant given Singapore's strict data sovereignty regulations for critical infrastructure.
The partnership between SMRT and Oracle is not new. Strides Technologies, SMRT's engineering arm, has previously collaborated with Oracle on other digital transformation initiatives. The Jarvis project, however, marks one of the most advanced applications of AI in rail maintenance in Southeast Asia. It builds on earlier experiments with predictive maintenance using IoT sensors on trains, but Jarvis goes a step further by unifying all data sources into a single AI-powered knowledge base.
Impact on workforce and operational efficiency
Ngien Hoon Ping emphasized that Jarvis is designed to augment, not replace, human workers. SMRT employs over 10,000 people, including engineers, technicians, and support staff. The AI tool handles repetitive tasks like data retrieval and initial fault diagnosis, freeing up skilled engineers to focus on complex problem-solving and strategic improvements. This is part of a broader upskilling initiative at SMRT, where staff are trained to work alongside AI systems.
Early results from the pilot phase show promising gains in productivity. Engineers using Jarvis report a 30% reduction in time spent searching for historical data, and a 25% faster diagnosis of recurring faults. The system also generates maintenance recommendations based on predictive models, which have already prevented several potential signalling failures during testing. Over time, SMRT expects these improvements to help it consistently meet the Land Transport Authority's MKBF target of one million train-kilometres.
Beyond maintenance, Jarvis is being explored for other applications, such as optimizing train schedules and managing power consumption. The platform's modular architecture allows new AI agents to be added as new data sources become available. SMRT is also considering integrating external data like weather forecasts and passenger flow patterns to further enhance predictive accuracy.
Broader context: AI in rail transport worldwide
SMRT's adoption of AI for predictive maintenance is part of a global trend. Rail operators in Japan, Europe, and the United States have been experimenting with similar systems. For example, East Japan Railway Company uses AI to monitor track conditions and predict equipment failures, while Germany's Deutsche Bahn employs machine learning to optimize maintenance schedules. However, SMRT's approach is notable for its use of generative AI and a unified data platform, which makes the system more accessible to non-experts through natural language interaction.
Singapore's rail network is relatively young compared to older systems in London or New York, but it faces unique challenges due to high passenger density and a tropical climate that accelerates wear and tear. The three-hour night maintenance window is particularly tight, as the network operates from around 5:30 am to midnight. Any delay in fault detection can lead to service disruptions the next day, affecting hundreds of thousands of commuters.
Jarvis aims to reduce the frequency and duration of such disruptions. By moving from corrective maintenance to predictive and prescriptive maintenance, SMRT hopes to achieve zero unscheduled downtime in the long term. The platform also supports root cause analysis by correlating faults with environmental conditions, usage patterns, and maintenance history.
Looking ahead, Ngien Hoon Ping envisions Jarvis evolving into a platform that can be shared with other rail operators in the region. Many Southeast Asian countries are expanding their urban rail networks, and they face similar challenges of maintaining reliability with limited maintenance windows. SMRT's open approach to sharing its AI models could accelerate digital transformation across the industry.
Oracle, for its part, sees the rail sector as a growth area for its cloud and AI services. The company is working with several other transport operators globally on predictive maintenance and digital twin projects. The success of Jarvis could serve as a reference case for these initiatives.
In summary, SMRT's Jarvis platform represents a significant step forward in the application of AI to critical infrastructure maintenance. By combining 30 years of operational data with modern AI techniques, it enables faster, more precise fault resolution and supports a culture of continuous improvement. As the system matures and expands to cover more aspects of rail operations, it has the potential to set a new standard for reliability in mass transit.
Source: ComputerWeekly.com News