Research

Frontier research. Production discipline.

Miniml’s leadership includes active academics — University of Edinburgh professors, an Alan Turing Institute Fellow, and an ELLIS Scholar. We work at the boundary of machine learning research and enterprise systems, then carry the methods that hold up under scrutiny into production. Research that ships.

Recent research from the team.

Technical work from Miniml researchers and collaborators — new methods, evaluations, and findings from frontier model development. Each entry links to a short summary on this site, with the full paper one click away.

Selected work — what it buys you
  • DeCoRe Why our retrieval systems don’t fabricate citations — decoding that suppresses hallucinated content before it reaches the user.
  • GRADA Hardening retrieval against poisoned documents — the checks we run before your RAG layer touches untrusted content.
  • MMLongBench When long context beats retrieval and when it doesn’t — measured, not guessed.
  • Q-Filters Cutting inference memory and serving cost without sacrificing accuracy in production.
  • Behavioral-shift auditing Knowing when a deployed model has quietly changed — the statistical test behind our drift monitoring.
  • FLARE Reasoning you can audit — traces that show why the system answered, not just what it answered.