Domains
AI media systems, realtime production workflows, virtual production, post-production automation, generative commerce and operational tooling.
Independent R&DKRD · Applied Research & Development
APPLIED SYSTEMS · PROJECTS · COLLABORATION
KRD is the professional and research identity of Guglielmo “KRD” Guarnera. KRD builds software where creative production becomes too complex for disconnected tools: AI generation pipelines, script intelligence, legacy automation, media-server workflows, delivery operations, production memory and operator decisions.
Explore what KRD builds, the contexts each system serves and its current stage of development. Further technical material is available on request for collaboration and review.
System map
AI media systems, realtime production workflows, virtual production, post-production automation, generative commerce and operational tooling.
Each project presents its purpose, operating context, development status and selected evidence.
Start a conversation about collaboration, integration or a closer technical review.
Selected work
Distinct systems for distinct operational realities, with project context, selected demonstrations and opportunities for collaboration.
AI production control layer for audiovisual pipelines: VFX, post-production, advertising, cinema, TV and web media workflows with traceable runs, assets and review.
Script-to-production spine for film, series, advertising and VFX planning: narrative material becomes scenes, shots, context packets and timeline structure.
Short-form audiovisual production pipeline for IP worlds, episodic web content and branded narrative systems with approvals, credits and organized delivery.
Provider adapters, job specs, provenance, asset flow and execution surfaces.
Script, scene, shot, context packet and decision-memory structures.
LED volume thinking, realtime workflows, media servers and calibration-aware tooling.
Operational wrappers around non-API software, OCR and deterministic UI control.
Relationships, lineage, states and production history as first-class software objects.
Fast prototypes designed to survive contact with real operators and production pressure.
Method
The useful part of applied AI is rarely the model call. The useful part is the surrounding system: context, provenance, asset state, operator intent, recovery paths, repeatability and auditability.
The same applies to virtual production and VFX pipelines. The tool matters, but the system around the tool decides whether work survives under pressure.
The work is organized around evidence, operators and durable production context.
Index