Murali Rama Krishnan

Objective: Develop a scalable e-ticket approval system for subcontractors to log in via biometric kiosks into MMHE offshore premises. The system processed high volumes of e-token entries daily and integrated with project-based work breakdown structures (WBS) for approval workflows.

Team Size: 3 developers

• Handled 100,000+ e-token entries per day, ensuring the system could process large volumes of data seamlessly and in real-time.
• Managed 500 approvals daily from the respective WBS (Work Breakdown Structure) associated with various projects, maintaining a high level of accuracy and security.
• Achieved real-time processing of biometric data and e-token requests, allowing for smooth and uninterrupted entry for subcontractors.
• Delivered 99.9% system availability, ensuring that the biometric kiosks and backend systems were reliable and available to users at all times.
• The system was designed to scale, supporting further increases in subcontractor activity and expanding to additional offshore premises.
• Implemented a Service-Oriented Architecture (SOA) and an event-driven architecture, ensuring that backend services invoked by Backbone.js and Underscore could efficiently process large transaction volumes.
• Developed an automated approval workflow integrated with the WBS, reducing manual errors and speeding up the approval process thus approval times were reduced by 40%, ensuring faster access for subcontractors

Objective: Develop a scalable ecommerce backend platform for an intra-global retail market place within the

agent and partners (intra-commercial space for internal clients).
Team Size: 20+ developers, engineers, and senior developers.

• The platform handled over 500,000 transactions per day during peak periods. Initial issues with transaction latency were addressed by optimising Kafka configurations and improving Postgres indexing, resulting in a 20% increase in processing speed.
• Scaling the backend platform to manage over 1 million concurrent users required robust load balancing and microservices design. Challenges in load distribution were resolved by fine-tuning microservice instances and deploying auto-scaling strategies, improving traffic handling by 30%.
• Managing vast volumes of customer and transaction data required both structured (Postgres) and unstructured (MongoDB) storage. Faced with slow query performance during high traffic, database optimizations improved data access times by 25%.
• SWIFT payment integration posed challenges in handling cross-border transactions and compliance. Implementing additional security layers and streamlining the payment process resolved these issues, reducing transaction failures by 15%.
• Cloud infrastructure costs were reduced by 25% through the use of automated scaling and resource optimization strategies, balancing cost with high availability during peak usage.

Objective: Develop an IoT device management and orchestration platform to control building operations such as room controls, lift management, and scheduling for improved efficiency. 

Team Size: 46+ developers and engineers (both onshore and offshore team)

• Managed over 10,000 connected IoT devices, including room controls, lifts, and HVAC systems, processing data streams with minimal latency (<50ms).
• Initial challenges with high-latency data processing and unreliable device synchronisation were resolved by optimising Golang’s concurrency model and Apache Kafka’s configurations, resulting in a 30% reduction in response time.
• Streamlining complex data flows from thousands of devices using Apache Nifi initially led to bottlenecks. This was mitigated by refining flow configurations, resulting in a 25% improvement in data throughput.
• Faced with scalability challenges due to growing IoT data, optimised cloud-native storage using Scylla for structured data and Redis for high-speed caching, achieving 99.95% data availability and a 40% improvement in query performance.
• Cloud resource costs were reduced by 20% through dynamic scaling strategies, balancing performance and cost by optimising the use of Scylla and Redis storage.

Objective: Develop a comprehensive CRM system for a large service-oriented enterprise to enhance

customer interaction, streamline processes, and integrate with core business operations. Team Size: 10+ developers and senior engineers.
Key Metrics & Challenges:

• Managed over 5 million customer records and handled up to 5,000+ daily customer interactions. Initial challenges included scaling the system to handle increasing data loads. This was resolved by optimising the MSSQL database and using Docker Swarm for dynamic resource allocation.
• Initial performance lags in the Blazor frontend led to a 20% drop in user satisfaction during the beta phase. Optimised gRPC-based server-to-client communication, improving responsiveness by 35% and regaining user confidence.
• Camunda integration for business process automation resulted in a 50% reduction in manual processes, cutting down service delivery times from days to hours.
• Established CI/CD pipelines with GitLab and Docker Swarm, reducing deployment times by 60% and ensuring quicker release cycles without compromising quality.
• Optimised cloud resources using Docker Swarm’s containerization, reducing cloud infrastructure costs by 25%.

Objective: Develop a robust system for managing real-time data processing platform projects, contracts, and

budgets, integrated with SAP and IBM Planning Analytics. Team Size: 15+ developers and senior developers.
Key Metrics & Challenges:

• Designed a system to manage 100+ ongoing projects simultaneously, handling a large volume of purchase orders and budgets totaling over $2+ billions.
• Initial challenges with handling the large influx of real-time data resulted in performance degradation. Resolved this by optimising Kafka configurations, increasing partition counts, and implementing dynamic scaling using MCC cloud infrastructure.
• Ensured 99.9% uptime by architecting the platform with a microservices architecture using Node.js and TypeScript, enabling fault tolerance and continuous system operation even during peak loads.
• Faced slow query performance while integrating with SAP. To resolve, optimised database schema and used MS-SQL stored procedures, which improved query performance by 30%. 
• Implemented CQRS pattern to isolate the command and query calls (especially by redirecting the correct write and read calls to the SQL server replica).
• Integration with SAP and IBM Planning Analytics required significant custom API development, resulting in initial delays. Through better cross-team collaboration, these were resolved in three months, enabling seamless data transfer and synchronisation.
• Reduced cloud costs by 20% through automated scaling mechanisms and optimising resource allocation in MCC cloud, particularly during off-peak hours.