Libardo Ramirez

Senior AI Systems & Backend Engineer

I build AI systems that ship — from architecture to production.

Open to consulting, contract & full-time

Curriculum Vitae

Senior AI Systems & Backend Engineer with 15+ years building software. Backend engineering, systems architecture, and cloud infrastructure, with recent work focused on production AI — RAG, local LLM inference, and on-device intelligence.

Remote / Hybrid · EU timezone

libar@libardoramirez.com LinkedIn GitHub

Experience

15+ years designing systems and leading teams

05/2024 – Present

AI Systems & Backend Engineer (Independent)

Remote, EU timezone

  • Design, build, ship, and operate production AI systems end to end — from architecture and APIs through deployment and ongoing operation
  • Built EQMO, a human-in-the-loop document approval platform for AI agents: a REST API and hosted MCP server with scoped keys, immutable SHA-256-hashed versioning, text-anchored review, and an append-only, hash-chained audit log
  • Ship privacy-first, on-device AI applications where document content and PII never leave the device, including an on-device RAG research workspace (EQBook) and AI-powered iOS apps built on Apple Foundation Models and MLX
  • Integrate local and open-weight LLMs (Ollama, MLX for Apple Silicon) into production, selecting and optimizing models for available RAM and latency budgets
  • Apply threat modeling, authentication/OAuth, sandboxing, and privacy-by-design across delivered systems, with secure secrets management in CI/CD and data layers in PostgreSQL
AI Systems RAG MCP Agents Local LLMs Python FastAPI Next.js AWS GCP Docker PostgreSQL
06/2021 – 05/2024

Senior Systems Architect / Senior Backend Engineer

PDPaola · SASMAT RETAIL S.L — Barcelona, Spain (Remote)

  • Led core system architecture, schema design, and third-party API integration strategies for high-volume enterprise e-commerce systems on a PHP stack
  • Architected asynchronous, queue-based service communication for the logistics integration (Logisfashion), replacing fragile synchronous calls and keeping order processing reliable under peak traffic
  • Designed token rotation and rate-limit-aware queue processing for partner APIs, eliminating recurring integration failures and removing manual recovery interventions from day-to-day operations
  • Standardized Docker-based deployments and environment configuration across all stages, cutting environment drift and shortening release cycles
System Architecture PHP Schema Design Queue-based Integration Docker CI/CD
2019 – 2021

MSc Studies, Information Technology Administration

Tecnológico de Monterrey

  • Completed postgraduate studies between international roles in Beijing and Barcelona
IT Administration Enterprise Architecture Technology Strategy
08/2014 – 06/2019

Systems Manager

Beijing International Bilingual Academy — Beijing, China

  • Directed hybrid infrastructure operations (on-premise Xen/KVM and cloud) with high availability and load balancing for critical academic and administrative systems
  • Led a 4-person cross-functional IT team covering support, network operations, and strategic deployments
  • Assessed and integrated open-source collaboration platforms to align with strict organizational data policies
Xen KVM Hybrid Infrastructure High Availability Team Leadership
12/2013 – 06/2014

IT Director

Gobe · SoftwareShop — Bogotá, Colombia

  • Established comprehensive IT strategy, aligning technical initiatives with organizational objectives during a critical transition period
  • Managed the technology budget and coordinated development efforts with a strong emphasis on security and operational reliability
IT Strategy Budgeting Security
06/2012 – 11/2013

Test Designer

Abits Colombia / AMADEUS Latam — Bogotá, Colombia

  • Designed and executed rigorous performance, stress, and security testing protocols on large-scale travel e-commerce platforms
  • Collaborated with engineering teams to identify and remediate critical security vulnerabilities before production release
Performance Testing Stress Testing Security Testing

Core Competencies

Organized by engineering domain

AI Systems

LLM engineering, RAG, embeddings and vector search, model evaluation, local and on-device inference (Ollama, MLX on Apple Silicon), multi-agent workflows with scoped permissions and output validation.

Backend

Python, FastAPI, Django, REST/GraphQL API design, PostgreSQL, MongoDB, schema design, queue-based and distributed service integration.

Infrastructure

AWS (EC2, S3, RDS), GCP, DigitalOcean, Docker, CI/CD, hybrid infrastructure and high availability, managed vs. self-hosted trade-offs.

Apple / On-device AI

Swift, SwiftUI, Apple Foundation Models, MLX, Core ML, SwiftData — building and shipping native, privacy-first applications.

Security

ISC² Certified in Cybersecurity (CC), threat modeling, zero-trust design, authentication & OAuth, sandboxing, secrets management, privacy by design.

Languages: Spanish — Native · English — C1 Fluent

Education & Certifications

ISC² Certified in Cybersecurity (CC)

ISC2 · 2025 — Threat modeling, zero-trust architecture, and vulnerability management

MSc — Information Technology Administration

Tecnológico de Monterrey, Mexico · 2019–2021 — Enterprise architecture, IT governance, and technology strategy

Information Security Specialist

Universidad de los Andes, Colombia · 2010 — Secure systems design and security operations

Systems Engineering and Computer Science

Universidad Industrial de Santander, Colombia · 2008 — Software engineering fundamentals and computer systems

Selected Work

Production software, shipped and available to users

EQMO Live

Human-in-the-loop document approval for AI agents. Agents submit over REST or a hosted MCP server; humans approve with text-anchored comments. Immutable, SHA-256-hashed versions and an append-only, hash-chained audit log.

Next.js REST / MCP Audit log

Bedtime Snuggles Live

AI children’s storytelling for iPhone and iPad, built on Apple Intelligence and Foundation Models. On-device generation, no accounts, no data collection. Live on the App Store.

SwiftUI Foundation Models On-device

EQBook Live

On-device AI research workspace for macOS: NotebookLM-style Q&A with embeddings, vector storage, and RAG inference running locally on Apple Silicon with MLX. Document content never leaves the device.

MLX RAG Local LLMs

EarnScreen Published

Privacy-first iOS app that gates distracting apps behind motion-verified real-world habits, using the Family Controls framework with fully on-device state.

Swift Family Controls Core Motion

Let's talk

Available now for remote or hybrid engagements in the EU timezone. Book a meeting or send an email.

Schedule a Meeting libar@libardoramirez.com