Principal Software Architect & Data Engineer
Specializing in High-Throughput Distributed Systems, Zero-Trust Cloud Infrastructure, and Resilient AI Pipelines.
flowchart LR
subgraph CLIENT ["1. Presentation Layer"]
UI["React 19 / Modern SPA"]
end
subgraph GATEWAY ["2. Edge DMZ Subnet (OCI)"]
NGINX["NGINX Reverse Proxy / TLS 1.3"]
SPRING["Spring Boot 3 Gateway (Java 17)"]
DB[("MySQL 8.0 / PostgreSQL")]
NGINX --> SPRING
SPRING --> DB
end
subgraph AI_ENGINE ["3. Private AI Subnet (OCI)"]
FASTAPI["FastAPI Inference Engine"]
ONNX["ONNX Runtime INT8 (Sub-18ms)"]
FASTAPI --> ONNX
end
UI -->|HTTPS| NGINX
SPRING -->|Private RPC| FASTAPI
style CLIENT fill:#0B0F19,stroke:#38BDF8,stroke-width:2px,color:#FFFFFF
style GATEWAY fill:#0B0F19,stroke:#34D399,stroke-width:2px,color:#FFFFFF
style AI_ENGINE fill:#0B0F19,stroke:#818CF8,stroke-width:2px,color:#FFFFFF
| Project | Core Domain & Stack | Live Demo & Artifacts | Highlights |
|---|---|---|---|
| Talent Acquisition AI Engine (Apply on Job) | Applied AI β’ Vector Search β’ DuckDB OLAP | π Live Web Demo | Two-stage AI matching (12k/s), Zero-Trust prompt firewall, NYC Law 144 compliance, 3-tier triage. |
| Data Sentinel | FinTech β’ XAI Fraud β’ AES-256-GCM | π Live Web Demo | Multi-tolerance reconciliation ( |
| Healthcare AI Anomaly Engine | Unsupervised ML β’ Plotly.js β’ Healthcare | π Live Web Demo | Isolation Forest multi-dimensional outlier detection on 50k+ Medicare claims (<3ms/row). |
| SaaS Churn Intelligence | Predictive ML β’ Survival Cohorts β’ Scikit | π Live Web Demo | SMOTE-balanced Logistic Regression (AUC 0.842), 24-month cohort survival curves protecting $1.42M ARR. |
| Aegis Stream | Streaming β’ Redpanda β’ ClickHouse OLAP | π Live Web Demo | Real-time SIEM event streaming (>50,000 evts/s) with ClickHouse MergeTree columnar partitioning. |
| TechMind | FastAPI β’ ONNX INT8 β’ Spring Boot β’ OCI | ποΈ Architecture Spec | Dual-Subnet OCI Cloud DMZ, local ONNX INT8 inference (<18ms), Groq LPU resilient cascading. |
| Logistics Analytics Hub | DuckDB β’ Kimball Star Schema β’ Power BI | π 25+ DAX Measures | In-memory OLAP data mart, C-Level Excel workbook, and sub-15ms spatial query views. |
| Principle | Architectural Implementation | Business & Operational Impact |
|---|---|---|
| Zero-Trust Network | All internal microservices, inference engines, and databases operate strictly within isolated private subnets with no public IPs. | Eliminates external attack surfaces; guarantees 100% compliance with enterprise security audits. |
| Local-First & Quantization | Execution of mathematical and NLP models locally on CPU using ONNX INT8 instead of relying on costly external APIs. | Decreases inference latency by 98% and eliminates recurring cloud GPU compute expenses. |
| Deterministic Data Lineage | Cryptographic SHA-256 content hashing for entity IDs instead of random UUIDs. | Eliminates duplicate ingestion, ensures exact idempotence, and provides mathematical auditability. |
- π Licenciatura en Ciberdefensa (In Progress) β Universidad de la Defensa Nacional (UNDEF)
- βοΈ Oracle Cloud Infrastructure Certified Foundations Associate β Oracle
[Verification ID: 103477615OCI26FNDCFA] - π‘οΈ CertificaciΓ³n Profesional en Ciberseguridad β IBM SkillsBuild
- π Data Science Tech Foundation & Machine Learning β Oracle
- πΎ Autonomous Database & Oracle APEX Specialist β Oracle Next Education / Alura Latam
- π IELTS Professional English Proficiency β Santander | British Council
β‘ Despliegues Activos: Mantenimiento de arquitecturas y pipelines en producciΓ³n.