Proposed capability · subject to evidence gates
Let machines execute bounded loops while people retain the objective, limits and stop authority.
The technical case remains rigorous while the public meaning stays clear and inspectable.
A proposed closed-loop laboratory could connect human-defined objectives, constrained experimental design, robotic execution, analytical measurement, provenance and model-guided selection of the next experiment.
Autonomy would be task-bounded. The system could optimise only inside an admitted variable, material and equipment domain; it could not redefine the scientific objective, expand the hazard envelope or make clinical, quality or release decisions.
A robot could repeat a carefully defined make-and-measure cycle and suggest the next test. Scientists would still choose the question, approve the limits and decide what the result means.
The useful test is not whether the robot looks futuristic. It is whether the workflow learns more safely and reliably than the real alternative.
Can a constrained robotic workflow reach a defined formulation or materials decision using fewer samples while preserving identity, repeatability and safe recovery?
Decision useAdopt one bounded automated step, redesign interfaces, retain the manual route or stop the automation case.
Expected improvement can choose a useful next experiment under declared constraints. It cannot know an unmeasured hazard or decide that an objective is scientifically or ethically valid.
xₙ₊₁ = argmaxₓ EI(x) subject to gⱼ(x) ≤ 0Every system would need justified demand, competent operators, utilities, safety controls, maintenance and an accountable intended use.
Move identified labware through approved physical states.
Create traceable candidates inside an approved mass and material range.
Challenge labware presence, orientation and identifier continuity.
Coordinate approved steps, versions, states and human gates.
Preserve sample, method, raw data and decision provenance.
The scientific workflow owner approves the objective and candidate space; safety and quality owners can pause, reject or retire the system independently.
The system cannot admit new materials, rewrite constraints, conceal exceptions, self-approve changes or make final scientific, clinical, quality or release decisions.
A comparative automation-feasibility record covering learning rate, repeatability, material use, operator burden, failure modes and safe recovery.
Stop when the manual baseline is safer or stronger, or when identity, provenance, interface reliability, human control or measurable value is insufficient.
Shared governance connects Glasgow and Hyderabad. Climate, infrastructure, demand and regional value keep their designs materially different.
The campus expression is a planning hypothesis only. It does not represent a secured site, approved design, funded programme, partner commitment or operating capability.
The campus expression is a planning hypothesis only. It does not represent a secured site, approved design, funded programme, partner commitment or operating capability.
Evidence that one repetitive workflow benefits from automation.
A bounded feasibility comparison against a manual baseline.
A realistic but controlled interoperability challenge.
An interface and recovery evidence package without product endorsement.
A governed route from algorithm to physical experiment.
A scoped closed-loop study with rights, limits and human gates explicit.
Future intent is not current performance. Every measure requires a defined workflow, boundary and accountable record.
Illustrative proposed experiment envelope. This is not performed work, a protocol, an installed capability, a service offer, a validated method or evidence of an operating laboratory. No GMP, GLP, manufacturing-licence, clinical or regulatory-readiness status is claimed.
Approved use case, objective and constraint set
Physical guarding, safe state and recovery drills
Identity and data-integrity challenge
Cybersecurity and supplier-dependency assessment
Independent human pause and override
Stage-appropriate software validation where the intended use requires it
These sources inform the proposed programme. They do not prove VARUNÉ capability, affiliation, performance, compliance or regulatory acceptance.
Demonstrates integration of robotics, computation, characterisation and active learning in a bounded materials workflow.
CaveatReported study performance belongs to that platform and does not prove general autonomy or pharmaceutical suitability.
Open primary sourceShows a mobile robot and Bayesian search across a defined ten-variable experimental space.
CaveatThe experiment is a research precedent, not evidence of a VARUNÉ platform or universal productivity.
Open primary sourceProvides current UK policy context for digital, automated and robotic medicines process development and manufacturing innovation.
CaveatA competition scope is not evidence of a VARUNÉ campus application or award, or that VARUNÉ has developed the technologies described. NEUVIOR's separate funded-project record is stated in the evidence hub.
Open primary sourceA conversation or published concept does not create a partnership, service, installed capability, programme commitment or authority to use another organisation’s name.