AI Factory & GPU Infrastructure
GPU clusters built to actually run: Nvidia DGX and GPU nodes on Kubernetes, scheduling, storage and networking sized for real workloads - designed, deployed and validated end to end.
Open to new engagements · Remote, worldwide
Razvan-Ioan Petrescu Infrastructure & AI Architect.
For 14+ years I have designed and run the infrastructure critical systems sit on - from AI Factory and GPU clusters to Kubernetes, AWS, GCP and on-prem estates, for banks, stock exchanges, industrial groups and HPC platforms. I design it, validate it, and leave it secure and documented. Now consulting independently from Alicante.
01 · Profile
Engineering quiet
Engineering quiet
infrastructure.
I help organisations move beyond fragile, hand-rolled systems toward infrastructure that is observable, automated, and provably secure - the kind of stack that simply does not wake people up at night.
My career has taken me from a stock exchange in Sibiu to a national security vendor, from industrial robots in the Carpathians to regulated financial institutions across Europe. The brief is always the same: make critical systems resilient, automated, and defensible - without slowing the people who depend on them.
I work end-to-end. Architecture diagrams to Ansible roles. Threat models to Grafana panels. AWS landing zones to bare-metal Kubernetes. I prefer evidence over opinion and small, durable improvements over heroic rewrites - and I hand off with runbooks and diagrams a successor could pick up cold, not tribal knowledge.
Today I operate as an independent consultant and fractional CTO serving banking, fintech, public sector and HPC clients across Europe - including AI and GPU platforms, and environments where GDPR, PCI-DSS, DORA and NIS2 are not theoretical. I speak at security conferences, mentor engineering teams, and write for openthreat.ro.
14+
Years in infrastructure
9
Sectors served
15+
Active certifications
4
Languages spoken
02 · Services
Consulting services.
I take on a small number of long-running engagements per year. Most start as one of these; all of them end with something more resilient than they started.
Infrastructure Architecture & Validation
A structured review of your infrastructure - what's fragile, what's exposed, what won't scale - followed by benchmarking, HA and capacity validation, with an acceptance report you can hand to a board. Typically 1–3 weeks.
Cloud Migration
Landing zones, workload migrations and cost/HA re-architecture across AWS and GCP, executed with minimal downtime and a tested rollback plan.
Kubernetes Consulting
Cluster builds, hardening and Helm-based delivery on EKS, GKE, RKE2 or bare metal - from first cluster to multi-cluster GitOps.
Fractional CTO / Technical Leadership
Senior technical leadership without a full-time hire - architecture sign-off, roadmap and vendor decisions, team mentoring, and vCISO-style security governance when you need it.
SIEM / SOAR Build-outs
Splunk, Wazuh and Elastic deployments tuned for real detections - not just log volume - plus the automation to act on them.
Security Audits & DORA / NIS2 Compliance
Infrastructure and web-app penetration testing mapped to the frameworks your auditors actually ask about - GDPR, PCI-DSS, DORA, NIS2.
Speaking & Workshops
Conference talks and internal workshops on SIEM, Kubernetes and infrastructure security, drawn from real production incidents.
03 · Trajectory
A decade plus
A decade plus
of mission-critical work.
From university IT to senior architecture roles in cybersecurity vendors - every step shaped by the discipline of running things that cannot afford to fail.
-
Founder & Principal Consultant
OpenThreat- Through my own firm I deliver the independent engagements summarised under Projects - AI Factory and GPU platforms, infrastructure architecture and validation, cloud, Kubernetes and security - for banking, financial-services and public-sector clients across the EU, all under NDA.
- Currently embedded as Senior Infrastructure Engineer (long-running contract) with a proxy and network-infrastructure provider - server and network infrastructure at scale, IaC automation, Kubernetes, observability and reliability, plus architecture and delivery of AI-as-a-service solutions.
- The same roof hosts the OpenThreat security conference series and openthreat.ro, where I publish field notes.
KubernetesAnsibleTerraform AWSNvidia DGXGrafana -
Senior Infrastructure Engineer
Bitdefender- Orchestrated full AWS → GCP migration of mission-critical systems, achieving HA and cost-efficiency.
- Designed and managed multi-cluster Kubernetes (on-prem + GCP) with auto-scaling and CI/CD.
- Built an internal solution for DNS subdomain takeover prevention, integrated with all major providers.
- Implemented WAF and host-based FIM on hybrid infra; integrated with Splunk, Wazuh, Elastic.
- Acted as technical lead and advisor for cloud infra, security posture and team mentoring.
AWSGCPKubernetes AnsibleTerraformSplunk WazuhElasticZabbix GrafanaGitLabNvidia DGX -
Senior Advisor, Network Systems Management
Secureworks- Managed enterprise-scale Splunk SIEM cluster; improved performance, retention, search and alerting.
- Developed detections for brute-force logins, impossible-travel events and EDR alerts.
- Automated data ingestion via Python and integrated third-party logs (EDR, WAF, DNS).
- Defined SIEM architecture best practices; led the Splunk team assigned to the client.
SplunkPythonLinux Security AutomationServiceNow -
Senior Infrastructure Engineer & Pentester
Zipper Services- Led the Linux team across research, design, implementation and operations.
- Architected and deployed secure CI/CD/CS pipelines using GitLab, Ansible Tower, Nessus Pro.
- Built monitoring with auto-deployed agents (Prometheus, Grafana, Ansible).
- Conducted internal pen-tests on infrastructure and web apps; produced reports and remediation pipelines.
- Built custom Dell EMC ECS auditing on Kibana + Elastic; SIEM with Windows / Linux audit and port mirroring.
GitLabAnsible TowerSplunk Nessus ProElasticVMware AlienVaultVeeam -
Infrastructure Engineer
Smart-X Net App- Maintained, configured and planned multi-datacenter infrastructure with Ansible Tower.
- Designed and implemented Vault + Consul cluster, Nginx-based WAF, HA VPN with global ACLs.
- Managed MongoDB and MySQL (master–relay–slave) plus DR planning and validation.
VaultConsulOpenStack OpenNebulaMongoDBMySQL OpenVPNIPSECCisco -
Infrastructure Engineer
Holzindustrie Schweighofer- Designed monitoring connecting datacenter sensors to industrial robots; managed backup and HA.
- Implemented secure video-monitoring and remote access via Cisco ASA VPN.
- Operated multi-vendor network and VMware clusters across an industrial campus.
VMwareVeeamCisco ASA ActiveDirectoryBitdefender -
System & Network Engineer
Sibex - Sibiu Stock Exchange- Configured Linux servers for stock-exchange applications; hardened the network against unauthorised access.
- Implemented MySQL master–slave architecture and NagVis availability maps.
LinuxNagiosiptables MySQLBGPCisco -
System & Network Administrator
Apulum Technologies- Built an on-prem mini-datacenter; managed AWS VPC, EC2, RDS, WAF, Route 53.
- Led the development team and implemented Nagios + Dynatrace observability.
-
System & Network Administrator
"1 December 1918" University of Alba Iulia- Planned and configured campus-wide network equipment, servers, and conferencing systems across all faculties.
04 · Capabilities
A toolkit built for scale.
Hands-on across the full stack - from kernel tuning to cloud landing zones - with a bias for automation, observability, and security by default.
AI & GPU Infrastructure
- Nvidia DGX
- GPU on Kubernetes
- HPC platforms
- Tanzu
- NSX-T
- vSphere / vSAN
- Capacity validation
Cloud & Platform
- AWS
- GCP
- OVH
- OpenStack
- Proxmox
- VMware
- vSAN
- Tanzu
Containers & Orchestration
- Kubernetes
- EKS
- GKE
- RKE2
- Helm
- Docker
- Docker Compose
- Istio
IaC & Automation
- Ansible
- Ansible Tower / AAP
- Terraform
- SaltStack
- Bash
- Python
- Ruby
Security & SIEM/SOAR
- Splunk
- Wazuh
- AlienVault
- Nessus Pro
- EDR
- WAF
- CloudFlare
- Hardening
Observability
- Grafana
- Prometheus
- Zabbix
- CheckMK
- Nagios
- OpenTelemetry
- Elastic / Kibana
DevSecOps Pipelines
- GitLab CI
- Bitbucket
- Helm
- Docker Hub
- Vault
- Consul
Linux Engineering
- Red Hat
- Ubuntu
- Debian
- Oracle Linux
- BSD
- AIX
- Kernel tuning
Networking & VPN
- BGP
- OpenVPN
- IPSEC
- L2TP
- Cisco
- Juniper
- Fortinet
- Nginx
05 · Independent practice
Selected private
Selected private
contracts.
A sample of the long-running engagements I have delivered as an independent consultant. All are under NDA and are described only at the level of scope and technology.
AWS Infrastructure & DevOps - financial services client
Long-running consultancy on a regulated AWS estate - landing zone hardening, GitOps with GitLab, Wazuh-based monitoring and serverless workloads.
AWSLambdaEC2RDSWAFCloudFormationIAMWazuh
L3 Technical Lead - banking client
Ran L3 infrastructure and security engineering across an Ansible Automation Platform estate, Kubernetes workloads, the API gateway and integration layer, and a VMware vSAN/vCenter virtualisation platform.
Ansible AAPKubernetesGrafanaEDRNessusVMwarevSAN
Senior Technical Consultant - banking client
Modernised the Linux estate and the automation backbone across Oracle Linux and RedHat, with a Bitbucket-driven Ansible workflow and hardened integrations between core platforms.
Oracle LinuxRedHatAnsibleBitbucket
Helm developer - public-sector client
Authored production-grade Helm charts and Kubernetes manifests for an internal platform spanning RKE2, RKE and Docker Compose footprints.
HelmKubernetesRKE2Docker
Security Consultant - public-sector client
Vulnerability management, SIEM and edge protection - Nessus Pro, Wazuh, EDR and Cloudflare WAF on a Debian/Ubuntu fleet.
Nessus ProWazuhEDRCloudFlare WAF
Presales Technical Consultant - GPU compute platform
Designed the compute platform for a multi-tenant, GPU-accelerated environment - Nvidia DGX, Kubernetes, NSX-T and the full VMware vSphere stack.
Nvidia DGXKubernetesNSX-TvSANvSphereTanzu
Senior DevOps Consultant - enterprise client
End-to-end DevOps and observability - Ansible-driven CI/CD, Kubernetes, Splunk, Grafana/Prometheus, Nessus Pro and Veeam-backed VMware.
AnsibleGitLabKubernetesSplunkGrafanaNessus
AWS DevOps - regulated product platform
Cloud architecture and DevOps automation for a regulated product on AWS - VPCs, RDS, S3, SNS and Kubernetes workloads.
AWSKubernetesAnsibleGrafanaPrometheus
Also · Open source
Tools I've shared publicly.
Small but useful - built once, shared so others can avoid the same yak-shave.
EC2 Backup
Bash script that snapshots and rotates EC2 volumes via the AWS CLI.
Graylog Log Generator
Python utility for stress-testing Graylog ingestion at realistic volumes.
Syslog Log Generator
Generates synthetic syslog streams for SIEM benchmarking and rule QA.
imapsync Web App
Web interface around the imapsync binary for safer mailbox migrations.
Ansible Roles
A small library of reusable roles for common Linux hardening and ops chores.
Docker Images
Curated images on Docker Hub for monitoring, log generation and ops tooling.
06 · Credentials
Verified certifications.
Continuously refreshed credentials across cloud, Kubernetes, Linux, security and emerging fields. Click any badge to verify on Credly.
Certified Kubernetes Administrator (CKA)
Linux Foundation · Jan 2026Linux Foundation Certified System Administrator (LFCS)
Linux Foundation · Nov 2025AWS Well-Architected Proficient
Amazon Web Services · Dec 2025AWS Knowledge: Amazon EKS
Amazon Web Services · Dec 2025AWS Knowledge: Data Migration
Amazon Web Services · Jan 2026Developing Helm Charts (SC104)
Linux Foundation · Jun 2025Introduction to Istio (LFS144)
Linux Foundation · Jun 2025Getting Started with OpenTelemetry (LFS148)
Linux Foundation · Jun 2025Quantum Computing Essentials (LFQ102)
Linux Foundation · Jun 2025EU Cyber Resilience Act (LFEL1001)
Linux Foundation · Apr 2025Splunk Enterprise Certified Admin
Splunk · 2021Splunk Core Certified Power User
Splunk · 2021NSE 2 - Network Security Associate
Fortinet · 2021 · ID mb8L8zdOE1NSE 1 - Network Security Associate
Fortinet · 2021 · ID LF-yb502ttchk→ Full credential history at credly.com/users/razvan-ioan-petrescu
07 · Field notes
Writing & speaking.
Field notes from real engagements - published on openthreat.ro and presented at industry meetups.
SIEM · Wazuh
Real-time security analytics: Grafana dashboard integration with Wazuh
openthreat.ro
Virtualisation
Cutting costs and regaining control: migrating from VMware to Proxmox VE
openthreat.ro
Patch management
Automated multi-OS patch management for air-gapped infrastructure
openthreat.ro
DevSecOps
CI/CD & continuous security with GitLab on-prem and Nessus Pro
openthreat.ro
2018
iptables rules to scale
Speaker - Security Espresso 0x0F. A working session on building maintainable, performant iptables rule sets for production Linux fleets.
securityespresso.org/0x0f2015–2018
OpenThreat.ro by UAB
Co-organiser of the OpenThreat security conference series at "1 December 1918" University of Alba Iulia - bringing practitioners and students together.
openthreat.ro
08 · Education
Foundations.
Computer Science training at one of Romania's oldest universities - followed by a Master's focused on advanced programming and databases.
Master's - Advanced Programming & Databases
"1 December 1918" University of Alba Iulia · Romania
2012 - 2014Bachelor's - Computer Science
"1 December 1918" University of Alba Iulia · Romania
2009 - 2012
09 · FAQ
Questions, answered.
The things prospective clients usually ask before an engagement starts.
What is an AI Factory, and what does building one involve?
An AI Factory is the compute platform your AI workloads run on: GPU nodes, scheduling, storage and networking sized to the job. I design, deploy and validate them - Nvidia DGX and GPU nodes on Kubernetes - and hand over infrastructure your team can operate, with the benchmarks to prove it performs.
What does a fractional CTO do?
A fractional CTO gives you senior technical leadership on a part-time contract: architecture sign-off, infrastructure roadmap, vendor and platform decisions, team mentoring, and security governance when you need it - without the cost of a full-time hire.
How do your engagements work?
I take on a small number of long-running engagements per year and work remotely across the EU in English, Romanian, Spanish and Italian. Most start with a 1–3 week architecture or validation review and continue as hands-on delivery or ongoing advisory.
What does infrastructure validation include?
Performance benchmarking, high-availability and failover testing, capacity planning and security review - delivered as an acceptance report with findings ranked by risk and a prioritised remediation path.
Which technologies do you work with?
Nvidia DGX and GPU infrastructure, Kubernetes (EKS, GKE, RKE2, bare metal), AWS and GCP, VMware vSphere/vSAN/Tanzu and Proxmox, with Terraform, Ansible and GitLab CI around them - plus Splunk, Wazuh and Elastic on the security side.
Do you work with companies in Romania and Spain?
Yes. I work remotely with clients across the EU - including Romania and Spain - and can run the engagement in English, Romanian, Spanish or Italian. I reply within one business day.
10 · Get in touch
Have a system that matters?
I take on a small number of long-running engagements per year - AI Factory and GPU platforms, infrastructure architecture and validation, cloud migrations, and fractional CTO advisory for in-house teams. Tell me what you are building and what you would like to be true a quarter from now.
→ I reply within one business day.