Connecting Scientific Research & Industry
Bridgelight

  • The Essentials
  • UX
  • UI
  • Data Visualisation
  • Search & Discovery
  • Information Architecture
  • Design System
  • Frontend

Bridgelight built The Big Connector, a semantic search platform that mapped, matched and visualised capability across research and industrial networks. I was the designer on a five person founding team.

The flagship use case was rail. Bridgelight became a key delivery partner to the UK rail industry supply chain, answering a question nobody could answer by hand: who does what, across thousands of suppliers. The same engine was then pointed at five further sector networks, covering advanced manufacturing, energy systems, renewables, industrial engineering and food manufacturing.

It sold to universities, to the Catapult centres set up to translate British research into industry, and to government. That gave the platform two very different users who had to meet inside one interface.

Key takeaways

  • Two audiences, one product. Scientific and academic users wanted depth: the taxonomy, the confidence score behind a match, the ability to interrogate why two organisations had been connected at all. Commercial users wanted a shortlist and a name to call. Neither would tolerate being handed the other's product, and the business could not afford to build both.
  • The cost of the old answer. Mapping a business network by hand ran to £200k and five months, and the picture went stale the moment it was delivered.
  • Nine bespoke builds, one product. Every client had been sold a variation of the platform, built to request. The work was to find the single product hiding inside them, so that a new client meant new data rather than a new build.
  • The best feature was the one nobody found. Multi term semantic search was the platform's real advantage over a keyword tool, but users defaulted to single keywords and got the weakest results the product could give. Surfacing that capability, rather than hiding it behind an empty search box, became a central design problem.
  • Shipped and earning. The platform reached v2.1, revenue generating, with repeat customers across industry, academia and government.

Customers put it better than a case study can: "a tool that makes us think about our work differently; not only beautiful, but very useful".

One journey, both audiences

Three screens, following a single rail manufacturer from a problem on the track to a supplier who can fix it. The academic user reads the same journey downward, into the evidence; the commercial user reads it forward, to the name and the number.

Research network screen: a graph of papers, researchers and manufacturers with the Rolling Contact Fatigue node selected, and a side panel showing evidence strength, manufacturing readiness, and three manufacturers ranked by match
01 — Research networkA graph on its own impresses people and helps nobody. Selecting a node has to answer a question, so the panel on the right turns a cluster of research into three facts a manufacturer can act on: how strong the evidence is, how close it is to being buildable, and who can build it. The academic reads the top of that panel, the commercial user reads the bottom, and neither is handed the other's product.
Manufacturing partners screen: suppliers ranked by match score against a brief, the top one showing a green panel of supporting evidence, the second showing an amber panel describing a gap
02 — Manufacturing partnersThe shortlist, and the answer to who does what across thousands of suppliers. A score alone is a black box, so every result carries its reasoning: the top match shows the evidence tying it to the research, the one below it shows the gap that costs six weeks of track possessions. Naming the weakness is what makes the ranking worth trusting.
Solution screen: a proposed engineering solution, a left to right chain from the manufacturer's problem through a paper, a researcher and a co-authored trial to the supplier, and a table checking each requirement against the supplier's capability
03 — Solution and evidenceThe screen that has to survive interrogation. The strip across the middle traces the connection the platform is claiming, link by link, from the problem on the track to the supplier at the end of the chain. Beneath it the requirements are checked one by one, and five of the six are met. Naming the sixth is the point: a match the user cannot audit is a match they will not act on.

Discuss project or contract work?

EMAIL ME

CALL ME

CV

Github

Linkedin

David Kinloch © Copyright 2026