Engineering at the center
Professional judgment leads. Technology gives it better context.
A NEW SPARK IN POWER ENGINEERING
Intelligence. Grounded in engineering.
We’re building a future where AI and software help engineers understand, diagnose, and maintain power infrastructure.
Born in Bolivia. Built with purpose.
Select a component to explore. An original anatomy study, not a specific equipment model.
01 OUR VISION
Better understanding.
Better-informed decisions.
The next step in maintenance starts with making sense of the evidence.
Power equipment tells a story through measurements, inspections, and technical records. We envision tools that connect that information, make knowledge easier to use, and support the engineer behind every decision.
Evidence before answers
Understand what the information supports, where it conflicts, and what is missing.
Knowledge that connects
Bring technical information into a clearer, more traceable maintenance workflow.
OUR FIELD OF FOCUS
- Transformers
- Autotransformers
- Power reactors
- Substation equipment
02 TECHNOLOGY & AI
Planned A thinking partner. An engineer in charge.
Our direction is to combine engineering software with advanced AI model APIs, including models such as Claude, to assist technical work. These integrations and capabilities are planned.
AI supports the engineer. Critical maintenance decisions remain under professional responsibility.
DOCUMENTS & KNOWLEDGE
Read the technical context
Assist with interpreting engineering documents, extracting relevant information, and organizing technical knowledge.
ANALYSIS & REASONING
Connect the evidence
Support technical analysis and compare diagnostic evidence, with explicit assumptions and limitations.
PLANNING & TRACEABILITY
Support the workflow
Assist maintenance planning and supervised technical and documentary tasks, with a record of the reasoning.
- Technical evidence
- AI-assisted analysis
- Engineering review
Human judgment at every critical step
Conceptual workflow. No live AI service is connected to this website.
03 FIRST RESEARCH PROJECT
Research stage CH / R–001
AGD Contraste
One sample. Multiple perspectives.
Exploring the interpretation of dissolved gas analysis (DGA) in mineral oil for transformers, autotransformers, and power reactors.
The goal is not a classification at any cost. It is to understand which interpretation the evidence supports—and what else we need to know.
- Compare diagnostic methods and experimental classification algorithms.
- Explore uncertainty, sensitivity, and inconclusive results.
- Trace the evidence, assumptions, and disagreements behind an interpretation.
When methods disagree, the evidence matters more.
Same evidence
Method A
Method B
Method C
Different interpretations
Look closer. Keep the uncertainty visible.
Conceptual visualization. No sample results or diagnostic outputs are shown.
TECHNICAL REFERENCES
IEC 60599:2022
IEEE C57.104-2019
Including applicable errata
References for research, not a claim of certification or conformity.
Research & prototype · Electrical Engineering thesis
AGD Contraste is linked to an Electrical Engineering thesis. Commercial evolution is envisaged after the thesis defense and is subject to subsequent validation.
04 RESEARCH & DEVELOPMENT
Rigor is part of the architecture.
A useful tool must make its reasoning inspectable. Our research direction puts limitations, uncertainty, and reproducibility in the foreground.
01
Compare
Contrast methods and experimentally evaluate classification algorithms without assuming a single interpretation is always sufficient.
02
Question
Study sensitivity to inputs, uncertainty, discrepant interpretations, and the boundaries of each method.
03
Document
Keep evidence and assumptions traceable. Identify inconclusive cases and the additional information needed.
An inconclusive result can be the most honest—and useful—result.
A PRINCIPLE THAT GUIDES OUR RESEARCH
05 FUTURE CAPABILITIES
A direction. Not a promise of shortcuts.
A conceptual roadmap for what we want to build. Beyond AGD Contraste research, the capabilities below are planned and are not commercially available.
01 Diagnostic contrast
Research stage
AGD Contraste: method comparison, experimental classification, and examination of uncertainty.
02 Technical knowledge assistant
Planned
AI-assisted reading, interpretation, and organization of engineering documentation.
03 Maintenance intelligence
Planned
Support for planning and managing maintenance processes, guided by engineers.
04 Supervised automation
Planned
Assistance with repetitive technical and documentary tasks, with review and traceability.
Development depends on research, technical validation, and responsible evaluation. No release dates are committed.
06 ABOUT CHISPITAS
A small spark. A long-term vision.
CHISPITAS began in Bolivia with a clear purpose: bring artificial intelligence and software engineering closer to the real challenges of power equipment diagnosis and maintenance.
We are an early-stage initiative. AGD Contraste is our starting point, within a broader ambition to build useful, explainable technology for electrical infrastructure.
BOLIVIA → WHAT COMES NEXT
THE PERSON BEHIND THE SPARK
CB.
Cristian Ballesteros
Founder & Project Lead
Cristian drives CHISPITAS and the initial development of AGD Contraste, with a vision of applying AI and software to power equipment diagnostics and maintenance processes.
Contact Cristian 07 LET’S CONNECT
Good ideas begin
with a conversation.
Interested in the intersection of power engineering, research, and AI? Get in touch.