In a bioinformatics classroom in Singapore, a cloud server can outlast the lesson. The students leave. The meter keeps running. In Relevance Lab’s account of a university deployment, the solution included end-of-day reports telling students and professors which servers were still on. Among the machinery of modern research, here was a remarkably ordinary problem: remembering to turn things off.
The useful version
- The job: make cloud infrastructure, enterprise data and AI usable inside real workflows.
- The method: engineers and managed services, backed by reusable software platforms.
- The lesson: put permissions, spending limits and approved tools where work begins.
That detail explains Relevance Lab better than a catalogue of technologies. Computing power is available. Making it accessible to the right person, under the right rules, takes another layer of work. This is where the company has placed itself: between the infrastructure an organization buys and the results it expects.
Put the rules inside the work
Research Gateway is the clearest expression of that idea. Universities, medical centers and life-sciences teams use research environments with particular tools, datasets and access restrictions. The platform gives them a self-service front door, with project budgets and governance attached. Researchers can select approved resources; administrators retain visibility into who is using them and what they cost.
An AWS technical walkthrough published in 2022 shows what sits behind that front door. A researcher can provision a genomics analysis environment, run a workflow and inspect its output. The architecture separates projects into AWS accounts. Its budget dashboard can track spending by project, researcher and product, with controls to pause or stop resources when thresholds are exceeded.
- 01ChooseApproved research tools
- 02RunProject-owned resources
- 03WatchUsage against budget
- 04ControlPause, stop, review
The Singapore classroom case adds student-level budgets and tools such as RStudio. An Australian academic medical-center example describes a different obstacle: on-premise capacity struggling with growing genomics data. Relevance Lab describes moving those workloads toward AWS Batch, keeping the deployment in Sydney and converting Singularity containers into Docker containers. A simple interface can conceal substantial engineering.
A services company with reusable parts
Relevance Lab sells that engineering as well as the interface. Its business covers cloud consulting, automation, product engineering and Data & AI. Customers can commission a migration, modernize an application, connect data systems or buy ongoing operational support. The reusable platforms give those engagements a starting point beyond a blank project plan.
RLCatalyst brings monitoring, service-desk integration and automation together. ServiceOne covers the daily business of infrastructure, application support, assets and security. SPECTRA addresses data integration, data lakes and warehouse modernization. These are connected needs: a useful analytics system depends on data arriving correctly, and a useful cloud application depends on someone keeping it running.
Lower operations cost in an unnamed publishing-company deployment described on the RLCatalyst product page. A case result, rather than a universal savings forecast.
The company’s displayed clients include Johnson & Johnson, American Textile and Analog Devices. Its narrower research specialty sits inside a wider enterprise-services business. In that market, buyers weigh implementation expertise and continuing support alongside software features. An internal platform team is also an alternative, particularly when an institution already has the skills to maintain its own environment.
The present organization reflects a combination of businesses. In January 2023, CIGNEX and Excellerent merged with Relevance Lab. CIGNEX brought open-source and cloud experience; Excellerent brought engineering capability and a development center in Ethiopia. The announcement put the combined workforce above 1,500. Today, the corporate site reports more than 1,550 people across a footprint spanning five countries.
The price of a simpler button
Research Gateway’s AWS Marketplace listing describes contract-based software entitlements. The cloud resources those users consume remain part of the economics. Compute, storage and data transfer need their own budget. Implementation and support depend on the arrangement. A procurement discussion therefore needs to separate the software agreement from the workload it will enable.
A historical MyResearchCloud article describes small research groups starting with monthly budgets of $100. Treat that as an illustration of a small workload, rather than today’s all-in price. The same article describes idle-resource stopping and budget enforcement. Its most transferable idea is operational: make expenditure visible while people can still change it.
There are conditions. An institution needs suitable identity controls, a workable data policy and tools that fit the environment. Sensitive research may require restricted data movement and reviewed exports. Moving a pipeline can mean changing containers and orchestration. Teams should test a representative workload and measure its complete cost before expanding access.
AI arrives at the same junction
In September 2026, Relevance Lab announced a majority investment in kaman.ai, an enterprise AI-agent platform. The stated purpose was to connect agents securely to existing applications and business workflows. It extends a familiar proposition: an impressive capability becomes valuable when it can operate inside an organization’s rules.
“Most enterprise AI efforts stall not on the model, but on integration”
Puneet Chaddah · Co-founder & CTO · September 2026
That is Chaddah’s diagnosis, and it gives buyers a practical question. Can the agent reach the information it needs, act with appropriate permissions and leave a reviewable record? Relevance Lab’s investment thesis puts those connections at the center. The announcement establishes a direction; individual deployments still have to establish their results.
A deadline makes the choice real
Another September announcement gave research customers a concrete decision. AWS said its 2026.09 release would be the final version of Research and Engineering Studio, with support ending September 30, 2027. Existing deployments can keep running. Maintenance and security responsibility will need a home. AWS lists Relevance Lab among partners supporting that transition.
For buyers, this is a useful way to evaluate the company: choose a real workflow, price its operation and decide who owns it after launch. The server left running after class is a small example. Multiply that question across laboratories and enterprises, and the work Relevance Lab sells becomes easy to recognize.