Applied AI · Financial-services client work / Client engagement
Turning early language-model capability into a useful scoring workflow.
Remote client delivery using React, Node.js and early GPT-3 capabilities for interview and question scoring.
The engagement
Through EnigmaScore, I contributed to a remote client engagement for Prospect 33. My CV documents React, Node.js and GPT-3 work for interview and question scoring.
What I delivered
The work joined frontend and backend delivery around an early language-model scoring capability, helping make the output usable within a practical application workflow.
Scope and evidence
The engagement was client work delivered through EnigmaScore. The public case study intentionally does not claim ownership of Prospect 33's products, private data, outcomes or current systems.
Why it matters
It was early hands-on experience applying a language model to a defined product problem: connect the input, model result and surrounding engineering workflow clearly enough for people to use the result.
Evidence & scope +
- The September 2026 CV records GPT-3 beta experience and remote Prospect 33 client work for interview/question scoring.
- Prospect 33's public website describes the company as working across AI, finance and production engineering.
- This card represents historical client delivery, not an assertion of ownership or a description of current Prospect 33 systems.