RCS // RESEARCH PROGRAM

Documented research. Transparent results.

An independent space for real technical questions, explicit methods and a strict distinction between measurements, external sources and estimates.

BACK TO PRESENTATION
RCS / METHODOLOGY

Publication rules

  1. 01Explicit question and method
  2. 02Visible data provenance
  3. 03Documented limitations
  4. 04Human validation before publication

01 // PROGRAM

Research projects

RCS-RP-001 and RCS-RP-002 are published with their reports and documented data. Future work remains explicitly labelled as such.

RCS-RP-001PUBLISHED

Supervised Multi-Agent Orchestration

RESEARCH QUESTION

How can specialised agents be assigned, coordinated and controlled while retaining traceable human validation?

Technical report v1.0 published on 18 August 2026. It establishes the feasibility of the initial six-role architecture and documents its limits; the live demonstrator has since evolved to request-scoped ephemeral specialists.

ORCHESTRATIONDESIGN SCIENCETRACEABILITYHUMAN CONTROL
DOI 10.5281/zenodo.21994886
RCS-RP-002PUBLISHED

Hardware-Aware Compression Performance Study

RESEARCH QUESTION

How do CPU, memory, operating system and hardware configuration affect compression algorithm performance?

Technical report v0.2 published on 26 August 2026. It combines a transparent 1,100-scenario model for 100 AMD and Intel CPUs, based on external sources, with a local 11-profile pilot on an Intel Core i7-13620H. Estimates and measurements remain distinct.

COMPRESSIONCPUMEMORYBENCHMARKS
DOI 10.5281/zenodo.22110548

RCS // EVIDENCE

Data classification

MEASURED

Produced directly by the described experimental protocol.

REPRODUCED

Obtained by repeating a documented protocol or test.

EXTERNAL

Imported from an identified third-party source.

ESTIMATED

Calculated by a model and never presented as a real measurement.

03 // PUBLICATION

Publish research

Restricted access. Enter the password to open the publication form.