Builder CV + GOTM ecosystem

Miroslav Šotek

Independent builder behind the God of the Math ecosystem: manuscripts, research engines, memory systems, answer- and action-control software, multi-agent coordination and containment, stochastic-computing neural work, control dynamics, quantum-control experiments, audio entrainment, public websites, and local infrastructure. Remanentia is one node in that portfolio, not a separate side project.

Contact ORCID GitHub

Profile

Research, systems engineering, and product work under one roof.

GOTM is the working archive and build environment behind ANULUM. It contains the theory, code, public sites, manuscripts, benchmark records, deployment mirrors, server work, product pages, and research traces needed to move ideas from private notes into systems that can be tested. The role is deliberately mixed: design the theory, write the code, run the systems, document the limits, and keep the public boundary honest.

1996-2026concept development horizon

2020-2026active code authorship horizon

GOTMsingle ecosystem, many surfaces

CH / LIANULUM operating base

Portfolio map

The entire GOTM ecosystem is the portfolio.

The complexity is intentional. Memory software, factual verification, stochastic neural computation, control research, manuscripts, civic publication, public websites, and server infrastructure all support the same larger work. Remanentia cannot stand alone if model output is unchecked. Factual control is weaker without memory. Control systems need simulation and measurement. Research products need public pages, licences, support paths, and operational infrastructure.

Evidence memory01

Remanentia

Remanentia is the memory and retrieval layer for the portfolio. It exists because long-running AI work fails when every session starts from a blank prompt, or when an answer cannot point back to the file, log, manuscript, benchmark, or decision record that supports it.

The current project state tracked in shared records is v0.5.0: a Rust-accelerated retrieval core, 3,003 tests, 60,890 vector chunks, compiled facts, MCP integration, device-aware indexing, PII redaction, and consolidation paths. Recent work added fleet-fed feed ingestion — admitting Finding and Decision records from the coordination bus through an admission gate — and cross-session entity synthesis. LongMemEval R11 is recorded at 72.2% overall, with temporal retrieval at 65.4% (oracle setting).

Its role in GOTM is not only storage. It is the evidence substrate for high-stakes LLM use: retrieval must preserve provenance, respect private/public boundaries, and make later review possible.

GitHub Account Benchmarks Platform

Answer & action control02

Director-AI and Director-Class AI

Director-AI is the answer-control layer. It reviews generated text against evidence, detects unsupported claims, and gives RAG or agent pipelines a checkpoint before output is trusted by a user, customer, researcher, or operator. It ships as the Apache-licensed director-ai package on PyPI, with a product page on ANULUM, a GitHub repository, and a Zenodo software record (Coherence Engine v1.4.1, published 2026-03-01).

Director-Class AI extends that work into the action plane. It is a separate commercial product (Business Source License 1.1) that sits between an autonomous agent and its effectors: it reviews high-impact shell, SQL, infrastructure, API, and MCP actions before they dispatch, escalates uncertain or high-risk actions to human approval, and emits tamper-evident audit records.

Together they close two gaps memory alone cannot: Remanentia can retrieve the evidence; Director-AI asks whether the generated answer stayed inside it; and Director-Class AI asks whether the action the agent is about to take is safe to run.

Director-AI Director-Class AI PyPI Zenodo DOI

Agent runtime03

Synapse Channel and HushLine

Synapse Channel is the coordination layer for multi-agent work: a local-first WebSocket hub where agents claim file scopes, share a plan, signal presence, message each other, and replay a durable event log after a restart — so several agents from any vendor cooperate across one repository or a whole ecosystem without colliding. Recent releases added declarative workflow orchestration (fan-out / map-join over the shared blackboard) and multi-host federation with mutual-TLS. It is AGPL-3.0, published as synapse-channel on PyPI.

HushLine is the containment layer: a local-only wrapper that filters, redacts, and bounds a command’s output before it reaches a prompt chain, with a per-directory execution permit and no network or telemetry. It is AGPL-3.0, published as hushline on PyPI.

With memory and verification, these complete a four-part agent runtime — remember, verify, connect, contain — and the whole fleet that builds the portfolio is coordinated through Synapse every day.

Synapse GitHub Synapse PyPI HushLine HushLine PyPI

Neural engine04

SC-NeuroCore

SC-NeuroCore is the stochastic-computing neural engine behind the hardware-facing side of the work. It treats efficient neural computation as a substrate problem: bitstreams, stochastic matrix operations, neuromorphic compatibility, and formal paths toward hardware validation.

The package is published as sc-neurocore-engine on PyPI and is active at v3.15.35. Local documentation tracks a large Python / Rust / HDL-facing codebase with a measured-fastest-first multi-language compute chain, and the Zenodo record describes SC-NeuroCore as a universal stochastic computing framework for neuromorphic hardware, published 2026-03-08.

Its role is strategic: if high-stakes AI depends on increasingly expensive dense computation, the portfolio also needs a route toward efficient, verifiable, hardware-aware neural systems.

Product GitHub PyPI Zenodo DOI

Control systems05

SCPN Control Family

The SCPN control family covers fusion-core, phase orchestration, classical control, and quantum-control experiments. The common question is whether coupled systems can be modelled, steered, measured, and reviewed without hiding the assumptions.

Local programme records identify these active code lines: SCPN Fusion Core v3.10.0, SCPN Control v0.22.0, SCPN Phase Orchestrator v0.10.0, and SCPN Quantum Control v0.10.0 (the last validated against IBM Heron r2 hardware). Public ANULUM pages expose the control, phase-orchestration, quantum-control, and wider SCPN surfaces.

This is the bridge from theory to intervention: solvers, benchmark scripts, controller comparisons, phase dynamics, and quantum-control experiments give the manuscript work something operational to test against.

SCPN Control Phase Orchestrator Quantum Control Fusion Core

Human state systems06

Fluctara

Fluctara is the audio-entrainment product line. It turns oscillator theory and SCPN-derived protocol ideas into user-facing sessions, API paths, and device-aware runtime work rather than leaving the concept only in manuscript form.

The engineering stack documented in the portfolio includes Rust-first computation, Python orchestration, protocol logic, EVS verification language, public API work, and a production site. It runs on its own infrastructure separate from Remanentia (a live frontend at fluctara.com and a backend health-checked API), and is at production grade. The pattern is shared: make theory measurable, inspectable, and usable.

The long-term value is a practical human-state system with clear limits: protocols, evidence, runtime metrics, payment/legal paths, and links back to the SCPN publication base.

Product site GitHub SCPN index DOI

Manuscripts07

God of the Math Corpus

God of the Math is the manuscript and research continuity behind the software. It contains the SCPN framework, layer papers, coherence-carrier work, control ideas, experimental notes, assumption tracking, extraction workflows, and the long-lived archive that makes the public projects part of one ecosystem.

Published Zenodo anchors currently include the SCPN master-publications index, the SCPN Framework v1.0 technical specification, and public Layer 1-3 records. The local corpus also contains wider manuscript families and validation packages that are not treated as published outputs until their public records are live.

This corpus is necessary because the code projects are not isolated utilities. They are implementations, tests, or product surfaces for a larger research programme.

Master index Spec v1.0 Publication list

Public websites08

Webmaster Layer

The webmaster layer is the public operating surface: product pages, documentation entrances, pricing, contact routes, legal pages, analytics, backup copies, FTP deployment, and domain-specific positioning. It turns private code and research into material that users, partners, reviewers, and funders can inspect.

It covers REMANENTIA-COM, ANULUM-LI, Fluctara, the Director-AI, Director-Class AI, Synapse, and HushLine product pages, service pages, SCPN project pages, and public mirrors. Product surfaces are kept current with each repository and carry honest, source-checked numbers rather than marketing claims.

This builder page is part of that layer: it explains the portfolio as a whole rather than presenting each product as an unrelated fragment.

ANULUM Remanentia Fluctara

Infrastructure09

Local Servers and Deployment

The infrastructure work covers local inference, workstations, the HP ProLiant ML350p Gen8 server, backup mirrors, benchmark machines, firmware and iLO maintenance, service runbooks, and the practical limits of running AI systems outside a pure cloud workflow.

Recent records show the ML350 on the local network with 2 × Xeon E5-2640, 128 GB ECC RAM, iLO 4 v2.82, CPU-only local LLM planning, and a first upgrade path toward 2 × E5-2697 v2. The server is useful for storage, indexing, RAG, embeddings, CI-style jobs, and small-to-medium local models before GPU expansion.

This matters because the portfolio has to run somewhere. Benchmarks, private corpora, memory indexes, backup discipline, and local inference capacity are part of the product, not an afterthought.

Infrastructure discussion Benchmarks

Public accountability10

Parazit.sk and Civic Work

Parazit.sk is the civic-publication line inside the wider portfolio. It is separate from the commercial AI products, but it shares the same concern with records, naming, evidence, and public presentation.

The reason it belongs on this CV is straightforward: the builder role is not only code production. It also includes publishing systems, public accountability, archive discipline, and the ability to present uncomfortable material in a structured way.

Within GOTM, this gives the portfolio a civic record function distinct from software, research manuscripts, and commercial product pages.

Parazit.sk

Zenodo and papers

Published records and citation anchors.

These are the public records currently safe to cite from the GOTM and ANULUM portfolio, split into the SCPN framework papers and the software records. The SCPN series is a larger, multi-paper research programme; only the specification and the published layer papers are listed as anchors here. Draft-only and metadata-generated records are intentionally not presented as published outputs.

Framework & theory

Public index

The SCPN Master Publications — Scope & Table of Contents

The public index for the God of the Math / SCPN publication set: the scope and table of contents for a layered model that treats consciousness and spacetime through phase-synchrony and informational transduction. Published 2025-10-22.

10.5281/zenodo.17419678

Specification

SCPN Framework v1.0 — Technical Specification

The reference architecture for the framework: its axioms, the layer interfaces, and the master Lagrangian that ties the multi-scale phase-synchrony and informational-transduction model together. Published 2026-02-06.

10.5281/zenodo.18507639

Paper 1

Layer 1 — The Quantum Biological Substrate

The framework’s lowest layer: the quantum-scale mechanisms proposed as the substrate from which higher-order biological coherence is argued to emerge. Published 2025-12-29.

10.5281/zenodo.18088340

Paper 2

Layer 2 — Neurochemical-Neurological Interface

Layer 2 as the transduction interface that links quantum-scale potentiality to neurochemical and neurological realisation. Published 2025-10-10.

10.5281/zenodo.17309834

Paper 3

Layer 3 — Genomic-Epigenomic Antenna

Layer 3 specifications, axioms, interface contracts, and the empirical programme: genomic and epigenomic coupling treated as a fractal antenna. Published 2025-11-02.

10.5281/zenodo.17508894

Software records

Software

Director-Class AI — Coherence Engine v1.4.1

The Zenodo software record for the factual-control / answer-verification engine line behind Director-AI: response-level evidence checking with a token-level streaming halt. Published 2026-03-01.

10.5281/zenodo.18822167

Software

SC-NeuroCore

A universal deterministic stochastic-computing framework for neuromorphic hardware design: bitstream-based neural computation with formal paths toward FPGA and hardware validation. Published 2026-03-08.

10.5281/zenodo.18906614

Software

SCPN Control

The neuro-symbolic control engine: Stochastic Petri Net → spiking-neural-network controllers, a Grad-Shafranov equilibrium solver, and multi-tier gyrokinetic transport. Software record for v0.22.0, 2026-06-26.

10.5281/zenodo.18821816

Software

SCPN Quantum Control

Quantum-native SCPN phase dynamics and control: NISQ simulation of coupled Kuramoto oscillators, with a first hardware observation campaign run on IBM Heron r2. Software record for v0.10.0, 2026-06-26.

10.5281/zenodo.18821929

Software

Synapse Channel

A local-first multi-agent coordination bus: work claims, a shared plan, presence, direct messaging, a durable event log, and declarative workflow orchestration. Software record for v0.71.0, 2026-06-29.

10.5281/zenodo.20801559

Software

HushLine

Local-only command-output containment: it filters, redacts, and bounds a command’s output before it reaches a prompt chain, with a per-directory execution permit. Software record for v0.1.5, 2026-06-20.

10.5281/zenodo.20775432

Software

SCPN-MIF-CORE

A magneto-inertial fusion core: a pulsed field-reversed-configuration (FRC) kinematic and RTL hot-path laboratory. Software record for v0.1.0, 2026-06-20.

10.5281/zenodo.20768029

Why it is needed

High-stakes AI cannot be only a model call.

When a system advises a person, answers a customer, summarises research, routes an operational task, or touches private records, fluency is not enough. The missing pieces are memory, source selection, verification, policy, latency measurement, and a way to inspect what happened after the fact.

That is the GOTM portfolio thesis: build the pieces that make AI systems less dependent on luck. Remanentia gives memory. Director Class AI gives factual control. The SCPN and SC-NeuroCore work explores alternative computation and control. The manuscript corpus keeps the long arc coherent. The public sites make the work legible enough for users, reviewers, partners, and funders.

1Remember the right source material instead of relying on a single prompt.

2Verify generated text against evidence before using it.

3Measure retrieval, latency, error modes, and deployment constraints.

4Turn the GOTM research ecosystem into services that people can evaluate and fund.

Functions

Expected functions across the portfolio.

Memory systems

Indexing, retrieval, vector search, compiled facts, operational recall, and public/private corpus boundaries.

Verification systems

Evidence checks, factual-consistency scoring, answer gating, refusal paths, and commercial licensing for closed deployments.

Compute systems

Rust acceleration, stochastic-computing models, WebGPU-facing work, and hardware-aware benchmarking.

Control systems

Phase dynamics, feedback loops, quantum-control experiments, and simulation surfaces for coupled systems.

Product systems

Documentation, pricing, support paths, deployment pages, legal pages, payment channels, and public project positioning.

Research systems

Benchmarks, reproducibility notes, manuscript work, peer-review channels, and long-term concept continuity.

Portfolio systems

Cross-project continuity, public mirrors, backups, session records, shared context, and the operational discipline needed to keep the whole ecosystem coherent.

Impact and potential

The near-term value is practical; the long-term value is structural.

For companiesSafer AI answers over private knowledge, clearer data boundaries, and deployment evidence that can be reviewed.

For researchersMemory benchmarks, factual-control methods, and control-system experiments that can be compared instead of only described.

For local operatorsSelf-hosted AI infrastructure that works with real constraints: hardware limits, corpus size, latency, privacy, and maintenance.

For the GOTM portfolioA single body of work whose pieces can support each other: memory improves verification, verification makes deployment safer, manuscripts preserve theory, and deployment reveals what research still has to solve.

Collaboration fit

Useful conversations are concrete: a memory failure, a retrieval problem, a factual-control requirement, a benchmark to reproduce, a hardware constraint, a research review, or a deployment that cannot afford unsupported answers.

Contact through Remanentia ANULUM contact page