Ever Allende@aandring671
1
Events attended
1
Submissions made
Argentina
10+ years of experience
About me
I am a cybersecurity auditor, full-stack builder and AI systems researcher focused on turning complex technical ideas into usable products. My work sits at the intersection of security, compliance, automation, blockchain infrastructure and applied AI. I build prototypes that are practical, auditable and designed for real-world deployment, not just demos. For this hackathon, I created RAE DecisionOps, an open-core trust layer for AI-assisted decisions. The goal is simple: help teams reduce overconfidence in noisy data environments and turn uncertain signals into clearer operational recommendations. I am especially interested in building AI-native infrastructure that can scale from local execution to cloud and GPU-backed workflows, while keeping reproducibility, security and business value at the center.
.png&w=640&q=75)
🤓 Latest Submissions
RAE DecisionOps AI for High-Stakes Decisions
RAE DecisionOps is an entropy-regulated trust layer for AI agents and business decision workflows. It helps teams convert noisy, incomplete or conflicting data into simple operational recommendations: act, observe or escalate. The system connects structured inputs, APIs, OCR outputs, market feeds, enterprise data and blockchain oracles through a modular pipeline. Before allowing a recommendation to be treated as reliable, RAE evaluates uncertainty, signal quality, confidence degradation, warnings and reproducibility metadata. The goal is not to build a magical predictor. The goal is to reduce overconfidence in weak-data environments and make AI-assisted decisions safer, explainable and audit-ready. The demo exposes a user-friendly interface where operators can define an Execution Context Matrix with variables, values, weights, confidence levels and notes. Business users receive a clear recommendation, while technical users can inspect regulated confidence, uncertainty, warnings and reproducibility hashes. For the AMD Developer Hackathon, RAE demonstrates how AMD infrastructure can support scalable AI-native workflows where performance, reproducibility and trust matter. The architecture is open-core: the public demo, technical walkthrough and reproducible runtime are available for validation, while production-grade decision matrices, scoring templates, enterprise connectors and deployment playbooks remain commercial modules. RAE can be applied to cybersecurity triage, compliance automation, market simulation, enterprise AI governance, agentic workflows and blockchain-connected data systems.
10 May 2026
