// about

About Me

Name Néstor Fleitas

Location Chubut, Argentina

Work mode Remote / Hybrid

📱 WhatsApp

// career path

Over 15 years in software engineering. I started as a Java backend developer working on SOA integration for banking and telecom (Banco Credicoop, HSBC, Telefónica, Telecom): distributed systems, service buses and integration architecture were my technical foundation.

That foundation evolved into cloud and operations: SRE at Equifax, mobile CI/CD at Flux IT, platform architecture at Ingenia, data & cloud engineering at Clarín, and DevSecOps at Allianz, where I worked from the security team on the corporate adoption of generative AI: secure development practices, access controls and LLM risk. Today I run DevOps platforms in banking (Banco Pichincha, Banco Itaú) with AWS, Azure, OpenShift, Terraform and GitLab CI.

That path converges in NexusOS: a governance platform for autonomous AI systems. I don't just build agents — I build the identity, authorization, policy, risk and audit layer that lets them operate safely in real environments. The core idea: AI proposes, governance decides.

📧 Send me an email CV in Spanish Resume in English

// expertise

Areas of Expertise

Three connected domains: platform and security experience is the foundation that makes it possible to design governed, production-ready autonomous AI systems.

A. AI Systems & Governance

Architecture of agentic systems and the control layer that makes them trustworthy.

B. DevOps, Platform & Cloud

Infrastructure and automation applied in banking, insurance, telcos and media.

C. Security & Software Engineering

Security across the software lifecycle, on top of a backend and architecture foundation.

// featured project

NexusOS

"AI proposes. Governance decides."

The problem

AI agents can already read databases, write to repositories and modify infrastructure, but most frameworks execute first and audit later: no audit trail, no approval workflow, no risk classification, and no identity model that answers who authorized what.

The approach

NexusOS is a governance and authorization layer for AI agents. Every capability goes through a pipeline of mandatory gates before execution: no shortcuts, no bypasses, and signed evidence for every decision. It is provider-neutral: governance is not rebuilt when you switch models.

// governance pipeline

LLM / Agent Identity Policy Permission Risk Verification Execution Audit

If any gate fails, execution is denied, the reason is logged and the audit record is still written. If the audit cannot be written, execution is denied as well.

✓ Implemented

Provider-neutral identity verification (Local, Keycloak, Azure Entra, LDAP)
Policy evaluation with signed evidence (HMAC-SHA256)
Risk classification: read-only / reversible write / irreversible write
Per-operator, capability-scoped permission grants
Append-only audit chain with prevHash linking
102 governed capabilities · 10 gates · 1,300+ test cases

⚡ In progress

Incremental execution enablement, starting with low-risk, reversible actions
Public-key signing for audit evidence (currently symmetric HMAC)

Use cases

Read-only infrastructure queries with operator identity and namespace scoping
AI-generated pull requests with a signed human approval before merge
Ticket automation and security operations with critical actions blocked until approval

Product & interactive demo ↗ GitHub ↗

// work

Projects & Case Studies

Governed integrations — Nexus Desktop & Senses

Agent integrations with Desktop, voice, Telegram, HTTP, Linux, Windows, Docker, Kubernetes, Git and logs, running under the NexusOS governance pipeline with multi-provider LLM routing.

Secure generative AI adoption

At Allianz Argentina, from the security team: assessment of secure development practices, access controls and LLM risk in the corporate rollout of generative AI; CI/CD security with Jenkins, SonarQube and GitHub Advanced Security.

Banking DevOps platforms

Infrastructure as code on Azure with Terraform and containers on OpenShift 4 (Banco Pichincha); platform administration and GitLab CI pipelines on AWS (Banco Itaú). CNCF framework evaluation in the DevOps chapter.

Data & Cloud in media

At Clarín (AGEA): GCP integration with Elastic Cloud (Pub/Sub, Dataflow, Airflow), Java connector development, EKS clusters with CDK and Node.js microservices.

// experience

Experience

// work history

2026
Present

Nexus AI — NexusOS

Founder & CTO · AI Systems Architect

Design and development of NexusOS, a governance and authorization platform for AI agents: a 10-gate pipeline (identity, policy, permission, risk, enforcement, audit), an HMAC-signed audit chain, and governed integrations for infrastructure, code and operations.

FEB 2026
Present

Banco Itaú

DevOps — Contractor

Administration of DevOps/DevSecOps infrastructure and platforms. CI/CD pipeline and deployment automation, infrastructure as code, and collaboration with technical teams on incidents and improvements. Stack: AWS, GitLab CI.

DIC 2024
FEB 2026

Banco Pichincha

DevOps — Freelance

Azure infrastructure with Terraform and container management on OpenShift 4. CNCF framework evaluation in the DevOps chapter and CI/CD process optimization in banking environments.

JUL 2024
MAY 2025

Allianz Argentina

DevSecOps

Worked from the security team on the corporate rollout of generative AI: assessment of secure development practices, access controls and LLM risk. CI/CD security with Jenkins, SonarQube and GHAS. AWS EKS infrastructure with Terraform. Security training for internal and external teams.

FEB 2024
JUL 2024

Arkho — Chile

Cloud Engineer — Contractor

AWS infrastructure with CloudFront, S3 and RDS. Backend with Node.js microservices on an EKS cluster. CDK, CodeBuild, CodePipeline, CodeCommit and Docker.

OCT 2022
MAR 2024

Clarín — AGEA

Data & Cloud Engineer

GCP integration with Elastic Cloud: Pub/Sub, Dataflow, Airflow. Docker and Kubernetes. Java connector development, AWS EKS with CDK (VPC, ELB) and Node.js microservices.

ABR 2022
SEP 2022

Ingenia

Technology Architect / SRE / DevSecOps

Geopagos: architecture design and AWS infrastructure with Kubernetes, Terraform, Airflow and GitLab. Mercantil Andina: new platform design on Azure with AKS, Terraform, Azure DevOps and Bitbucket.

JUL 2021
ABR 2022

Flux IT

DevOps Mobile

CI/CD for Android and iOS apps with Docker, Jenkins and Azure DevOps across GCP, AWS and Azure. Integration with SonarQube and Browserstack.

FEB 2020
JUN 2021

Equifax

SRE — DevOps

Site Reliability Engineering: Docker, Ansible, Apache, Tomcat, Nagios, Red Hat Linux, Nginx, Jenkins, Terraform and Kubernetes on GCP.

2012
2022

Java / SOA era — banking & telcos

Backend & Integration Engineer · Architecture

Backend and integration architecture foundation: Telecom (SOA with OSB 12c, DataPower, Jenkins; technical lead and configuration manager 2018–2020, plus production deployment support off-hours and on weekends through late 2022), HSBC (Sr Java, IBM ESB, 2017–2018), Banco Credicoop (Java in the Architecture team; open source research: CouchDB, Node.js, ELK, Jenkins, 2014–2017), Telefónica/Movistar (Java SOA/OSB consultant, WebLogic, Spring, 2012–2014).

2008
2009

Accenture

Cobol Developer

Cobol programming on mainframe (JCL, DB2).

// education & certifications

AWS Cloud Practitioner

Amazon Web Services

Red Team · OSINT · DevSecOps & Cloud Security

Hackademy (2022–2025)

CyberOps

SeaCCNA (2024)

DevOps · Data Science

MundoSE (2022–2023)

2013

Industrial Engineering

U.T.N. — Facultad Regional Buenos Aires

2002

IT Technician

E.E.T. N° 9 Ing. Torcuato Di Tella — Avellaneda

// services

Services

I work with professionals, startups and small businesses losing hours to manual work, and with platform and security teams that need to adopt AI without losing control.

Process automation

I map the manual, repetitive work, decide what is worth automating and implement it. The goal is concrete: fewer operational hours, fewer data-entry errors and a process that holds up without depending on one person.

Applied AI agents

Agents and assistants that do real work on your data and systems: answering questions, classifying, drafting, extracting information from documents. With defined execution boundaries and human review where the action is critical.

Tool and system integration

I connect what you already use — spreadsheets, email, WhatsApp, CRM, ERP, internal APIs — so information flows without copy-paste. Integrating well is almost always cheaper than buying another tool.

AI adoption with control

AI applied to a measurable problem, not to a demo: what data goes in, what the model decides, what stays audited. Includes secure development, prompt injection mitigation and human approval on sensitive actions.

DevOps / DevSecOps

So that what you automate is sustainable: CI/CD, infrastructure as code on AWS, Azure and GCP, Kubernetes, Terraform and security integrated into the lifecycle. Applied experience in banking and insurance.

Architecture review & technical advisory

An external senior perspective before committing investment: architecture review and guidance on platform, security and AI decisions.

Start a conversation Automation for small businesses: free assessment →

AI Systems Architect& DevSecOps Engineer