I design prompt systems, AI workflows, and instruction structures for practical applied AI.

Gustavo Lima  • Peerlist

VISUS is a photographic and visual AI system built to analyze black-and-white images and support LUT generation from their tonal structure.The system starts from the image itself. It evaluates luminance, contrast, tonal masses, regions, visual relationships, and validation points to treat the image as technical material, not only as visual reference.Its purpose is to transform a black-and-white reference image into a usable tonal profile for LUT generation. VISUS connects visual perception, measurable image structure, and final black-and-white output in a single applied AI workflow.

VISUS

ARGENTUM is a photographic AI workflow designed to turn analog film manuals and technical references into a structured process for creating black-and-white LUTs.The system reads manuals, identifies relevant data, structures tonal references, checks information against source material, and guides LUT generation with traceability. It connects photography, technical documentation, analog film behavior, and contemporary digital workflows.As a project, ARGENTUM demonstrates applied prompt architecture beyond text: a system that connects documents, tonal interpretation, validation, and final visual output.

ARGENTUM

FORTIA is an applied AI system designed for tax context, document organization, official-source research, and assisted guidance for Brazilian income tax situations.It was built for a sensitive domain where AI needs clear limits. The system works with source validation, traceability, risk awareness, and operational clarity. Its role is not to replace an accountant or provide definitive tax advice, but to structure more careful AI-assisted use in a high-precision context.As a project, FORTIA demonstrates responsible applied AI: when the context requires precision, the system must reduce assumptions, separate fact from inference, and make the limits of the answer explicit.

FORTIA

QUOTUS is an applied AI workflow for organizing quotes, suppliers, purchase lists, replies, and price comparisons.It was built to structure an operational routine that is usually scattered across spreadsheets, conversations, supplier contacts, manual lists, and isolated decisions. The system turns that process into a clearer workflow for comparing items, suppliers, prices, responses, and decision criteria.As a project, QUOTUS shows a direct application of prompt engineering to a concrete operational need: turning dispersed communication and supplier data into a more structured decision flow.

QUOTUS

ORION is a strategic AI system designed to support professional direction, editorial planning, decision-making, and execution.It organizes active goals, current priorities, future actions, operational constraints, and historical context without treating each conversation as isolated. The system helps turn scattered information into structured direction, keeping strategy, communication, career positioning, and execution aligned.As a project, ORION represents applied AI used as a strategic operating layer: preserving context, clarifying priorities, and supporting better decisions over time.

ORION

IMAGO is a critical writing system designed for analysis, review writing, and preservation of authorial voice.It works with a living corpus of previous texts to understand vocabulary, rhythm, structure, tone, and critical judgment before drafting or revising new writing. The system is built to avoid generic paraphrase, superficial praise, empty hooks, and stylistic dilution.As a project, IMAGO explores how applied AI can support writing without replacing voice. Its purpose is to combine external research, critical analysis, and authorial consistency into reviews that remain useful, specific, and recognizable.

IMAGO

MODULA is a technical system for evaluating, testing, refining, and releasing prompt systems with clear operational criteria.It was built to separate drafting from validation. After a prompt or prompt system is created, MODULA analyzes its structure, identifies gaps, evaluates consistency, supports controlled compression, tests weak points, and determines whether the system is ready for use, should be revised, or should remain suspended.As a project, MODULA represents a quality-control layer for prompt engineering: not a shortcut for creating prompts, but a structured way to validate their reliability, clarity, and readiness for real use.

MODULA

WERKON is a prompt systems and instruction architecture project designed to create, analyze, refine, and evolve prompts for professional AI use.It treats prompts as operational structures, not isolated commands. The system helps diagnose the real problem, define criteria, organize scope, correct failures, validate outputs, and adapt instructions to context.WERKON was built to turn vague or poorly structured requests into clearer, more reliable AI workflows for practical use.