RK Gokul

His engineering foundation was built in signal processing rather than software delivery. As a project trainee at a national satellite research centre he worked on filter design and development for satellite communications, running fixed- and floating-point simulation, memory management, and VHDL code generation verified against an industry simulator, before synthesising and testing the designs in real time on an FPGA kit and a DSP processor.

A parallel research placement covered physical-layer WiMAX design in simulation, FM receiver and OFDM transmit/receive models in a software-defined radio toolchain, and their realisation on SDR hardware tested against real-time signal generators and spectrum analysers, alongside embedded C for robotics equipment.

This is the part of his background that shapes how he approaches performance work today: a habit of measuring against a physical reality rather than trusting a dashboard, formed before he ever profiled an application.

He moved into performance testing and immediately took on system types that standard tooling did not handle well. He devised and led the implementation of an approach for performance testing content-delivery-network-based systems, where conventional load models give misleading results.

In a chemicals engagement he managed and executed performance tests for more than forty applications in a compressed timeframe, running multiple rounds across several environments. For a banking client he ran a proof of concept for applications using card-based authentication inside a highly secured environment.

Within the performance practice he built custom plug-ins for an open-source load testing tool using Eclipse and NetBeans, and provided expert analysis of database Automatic Workload Repository reports for database performance engineering — the beginning of the deep-dive diagnostic work that would define the next stage.

As a member of a performance Centre of Excellence he contributed directly to revenue, developing and presenting winning proposals and executing the high-stakes proofs of concept that decided them.

He led and mentored cross-functional teams of performance engineers across multiple projects and global locations, running bootcamp training and workshops on the major monitoring and load testing platforms. He also managed rapid consulting engagements across retail, banking and money transfer — defining a test plan, executing load, stress and spike tests, and giving stakeholders an answer inside a short window.

Two pieces of work from this period show the pattern. For a trading platform he produced a custom load generation tool in Linux shell and C/C++ because no commercial tool could simulate the specialised production load. For an investment banking client he developed scripts able to handle non-contemporary asynchronous calls over WebSocket connections. On a retail engagement he repaired ten-plus existing scripts while preserving flow integrity, saving 80% of the expected effort.

This stage was about removing effort and finding causes. By building out load testing infrastructure and developing reusable C and C# libraries and utility functions, he achieved a 90% reduction in test execution effort.

He introduced usage-pattern analysis built from production logs and application monitoring data, so that test scenarios emulated what real users actually did rather than what a specification assumed. Against that baseline he identified and resolved showstopper bottlenecks at both application and database level, using diagnostic tooling — database advisors, heap and thread dump analysers — to deliver root cause analysis and optimisation recommendations that development teams could act on.

He also extended what could be tested at all: desktop applications, thick-client Java systems and HTML5/WebSocket platforms, writing custom scripts and development-environment add-ins where commercial tools fell short. Two outcomes stand out — a travel and transport engagement where automating test data setup cut project timelines and effort by 85%, and a money transfer client where new script migration practice reduced version-upgrade effort by 80%.

Based in Stockholm, he led multi-domain consulting for a global retail client and drove the renewal and extension of contracts into two large additional domains on the strength of demonstrated value.

He architected and owned the performance environment and solution architecture for large-scale enterprise applications, working for scalability, simplicity and sound cloud configuration in Azure. By introducing automation, process change and targeted proofs of concept, he lifted the deliverable capacity of SAFe-Agile product teams by 30 to 50 per cent.

The result that matters most from this period is a negative number: by establishing enterprise-level quality standards and moving performance engineering into the CI/CD pipeline, his programmes reached zero production leakage of performance defects and escalations fell sharply. Senior client leadership described him as a "bright spot" and cited his approach as an example for other teams.

He served on governance and oversight committees, partnering with enterprise architects to define future-state architecture and present business cases for new technology to senior leaders, and championed Azure cost optimisation through detailed infrastructure analysis tied to business goals.

In his current work he established the performance engineering function across a complex enterprise application landscape — enterprise resource planning, analytics and high-volume e-commerce platforms — covering testing, deep-dive monitoring and proactive troubleshooting.

He has led several major cloud migration programmes under a zero-downtime mandate, including moving enterprise resource planning from on-premise to cloud, with performance held flat through the cut-over.

Over the past two years he has built Site Reliability Engineering practice around containerisation: proving the approach, then leading the containerisation of critical applications on Kubernetes and managed Kubernetes services to improve scalability and resilience. Optimising those deployments and the infrastructure underneath removed 50% of environment cost. The automation pipelines he designed for infrastructure provisioning and application configuration took eight hours out of every deployment cycle.

He acts as the definitive subject-matter expert on performance engineering and containerisation for development, operations and leadership teams.