QTrainer / Research software & delivery
The research needed software that implemented the methods faithfully, ran fast enough for live experiments, and could be used by the people doing the work.
I built QTrainer's RAG system and a self-composing knowledge graph connecting scientific specifications to implementation.
The platform is a low-latency, statistically validated research system using proprietary Bayesian updating and advanced statistical processing.
I built the servers, shipped them to a stakeholder, and handled onboarding while keeping the existing services running. The work covered the software, the hardware, and helping someone else operate it. QCC source and methods remain private. Independent retrieval demo ↗