Scientific research today relies heavily on cloud-based computing, HPC clusters, GPU workloads, and data-intensive workflows. Yet for many researchers, accessing these resources still feels slow, complex, and overly technical. Navigating cloud choices, interpreting costs, understanding policies, and configuring the right environment often becomes a barrier rather than an enabler.
Relevance Lab’s Research Assistant (RA) copilot solves this by adding a GenAI-native, conversational intelligence layer on top of the Research Gateway SaaS platform, which dramatically simplifies how researchers interact with cloud and HPC resources.
Why a Research Assistant is Needed
Cloud & HPC complexity is overwhelming
Choosing the right instance, storage type, GPU, or SLURM configuration requires cloud expertise that most researchers don’t have. For new students or collaborators, this barrier is even higher.
Research IT is overloaded
IT teams spend significant time answering repetitive questions such as workspace selection, job failures, dataset mounting, cost explanations. This slows down both IT productivity and researcher progress.
PIs need insights, not dashboards
Dashboards show data but not meaning. PIs want clear summaries of risks, trends, anomalies, and recommended actions.
Policies are hard to interpret
Institutions have strict guidelines for data governance and cloud usage. But researchers don’t know where to look or how to apply them.
This is where the RA copilot becomes essential. It adds an intelligent, policy-aware conversational layer on top of Research Gateway, turning complex cloud and HPC requirements into guided, compliant, and ready-to-execute actions.
How RA Copilot Works
The RA copilot converts high-level natural-language requests into precise, policy-compliant, and cost-aware actions.
A simple prompt such as “Create a GPU workspace for Cryo-EM image classification” is interpreted, clarified, and mapped to the right configuration.
The copilot handles intent understanding, resource recommendations, cost and risk insights, and policy validation. Once the request is fully resolved, Research Gateway executes the provisioning securely and consistently.
In essence, RA delivers the intelligence, and Research Gateway delivers the governed execution.
The RA Layer Handles:
- Understanding user intent
- Asking clarifying questions to remove ambiguity
- Mapping requirements to catalog items
- Recommending the most optimal compute, storage, or cluster setup
- Summarizing cost, usage, risks, and governance factors
- Answering university or institutional policy questions using RAG (retrieval augmented generation)
Research Gateway Handles:
- Provisioning secure workspaces
- IAM, policies, approvals
- Budget enforcement and cost controls
- Audit logs, dashboards, and integration with cloud platforms
Together, the RA copilot and Research Gateway deliver a guided, compliant, and ready-to-execute experience that simplifies cloud and HPC access. Researchers get the intelligence they need, and institutions get the governance they expect.
Key Capabilities of RA Copilot
RA ships with six high-impact capabilities:
- Create Compute – Provision workspaces and environments with simple prompts.
- Product Recommender – Suggests exact-fit or partial-fit catalog items.
- HPC Planner – Designs HPC setups including SLURM, FSx, EFA, and node sizing.
- Daily Operational Summary – FinOps, SecOps, and CloudOps insights delivered automatically.
- Security & SOP Assistant – Answers policy and compliance questions instantly.
- Extensible Actions – Add new workflows easily via GenAI + Tools + APIs.
Why It Matters
A New Blueprint for Research Computing
The RA copilot demonstrates how SaaS platforms can evolve using:
- Domain-tuned GenAI
- Institutional RAG
- Tool-based execution with guardrails
- Conversational workflows
The result is a powerful new operating model where: Researchers ask. RA reasons. Research Gateway executes.
This model establishes a modern foundation for research computing by combining GenAI-driven intelligence with secure, compliant, and automated execution. It streamlines access to cloud and HPC resources, reduces operational friction, and provides researchers with a smarter, more intuitive path to discovery.
Request a demo to see RA Copilot in action and start your journey toward modern research computing.