Gary Climacosa
Cloud Platforms,
Distributed Systems & AI Solutions
I work across AWS, Azure, and Kubernetes, with a recent focus on AI systems and hybrid architectures. Most of my portfolio work is hands-on: architecture case studies, working prototypes, and production-inspired systems built to test real constraints rather than just describe ideas.
- AWS and Azure platform delivery
- Kubernetes and distributed system patterns
- AI systems and hybrid architectures
- Real builds with visible constraints and tradeoffs
- FinOps and cloud cost engineering
What I Do
What I Build
The kind of platform, cloud, and AI work I tend to do best.
Cloud Platforms
I work on AWS and Azure environments where delivery, change control, and day-2 operations all matter. The focus is practical platform work: infrastructure as code, repeatable deployments, and clearer paths for teams using the platform.
"Automation should reduce risk, not add mystery."
Kubernetes & Distributed Systems
I use Kubernetes where it helps standardize deployment and operations across environments. The interesting part is usually not the cluster itself, but how services, data, and ownership fit together once the system is running.
"Use distributed systems deliberately."
FinOps & Cost Engineering
I bring FinOps discipline to infrastructure work: spend visibility, right-sizing, rollback paths, and operational clarity. The goal is not a clever stack — it is a system where cost drivers are visible and another engineer can support it without guessing.
"Make cost drivers visible. Choose the right envelope."
Practical AI Solutions
My recent focus is AI systems that stay close to real operational constraints: where data is stored, what can cross a boundary, how inference is wired in, and what the system costs to run. That work includes private retrieval systems, Bedrock-based prototypes, and hybrid AI setups.
"AI work starts with system boundaries."
Hybrid Delivery
Some of the most useful systems sit between environments, not inside a single cloud. I build and document patterns that connect cloud services, self-hosted components, and internal data in a way that stays inspectable and supportable.
"Hybrid systems need clear ownership and boundaries."
Selected Work
Projects, Prototypes & Case Studies
Some of these are architecture case studies, some are working prototypes, and some are production-inspired systems used to test how the pieces fit together.
EKS Triage Copilot
Working prototype
Problem
Kubernetes incidents often start with scattered context across logs, alerts, and cluster state, which slows down first-pass diagnosis.
Approach
Built a small serverless flow on AWS using Bedrock, Lambda, and API Gateway to turn raw incident context into a structured first-pass summary.
Outcome
A deployable prototype that shows how AI-assisted triage could fit into a Kubernetes support workflow, with very low running cost.
AWS Cost Dashboard
FinOps case study — 78% cost reduction
Problem
Engineering teams need actionable spend insights, not raw billing data — and the infrastructure itself should demonstrate cost awareness.
Approach
Built two deployment versions: V1 uses ECS Fargate + ALB for enterprise-grade delivery (~$27/mo). V2 swaps to EC2 t4g.nano + Caddy for ~$6/mo — same dashboard, 78% cost reduction. Both share the same Lambda + Athena data pipeline and GitLab CI lifecycle.
Outcome
A deployed, data-driven FinOps prototype demonstrating intentional infrastructure lifecycle management — including the ability to choose the right cost envelope for the workload.
OpenRAG Knowledge Platform
Production-inspired system
Problem
Teams want AI-assisted document search without pushing sensitive material into a broad public SaaS workflow.
Approach
Set up a private RAG stack with vector search, reranking, and cited answers on self-hosted infrastructure, while using cloud LLM services selectively.
Outcome
A working environment for testing private knowledge workflows, data handling choices, and the operational shape of a self-hosted AI system.
Homelab Architecture
Architecture case study
Problem
Portfolio systems are hard to evaluate if they only exist as diagrams or code snippets.
Approach
Built a small self-hosted platform with Proxmox, Kubernetes, storage, and internal tooling so projects can be deployed and operated end to end.
Outcome
A repeatable environment for testing platform ideas and showing how the moving parts connect in practice.
FrostStream CDC Pipeline
Production-inspired system
Problem
CDC pipelines need reliable event flow and usable storage layout without putting unnecessary pressure on the source database.
Approach
Built a pipeline around Debezium, Kafka, and MinIO to capture WAL changes and land them as partitioned files for downstream use.
Outcome
A working pipeline for exploring CDC design, failure handling, and downstream storage patterns in a contained environment.
Career Context
Selected Experience
Enterprise environments where I've delivered cloud, platform, and infrastructure outcomes.
TOPdesk Nederland BV
Cloud / SaaS Operations
- Platform support across cloud and SaaS environments with structured change and release coordination
- Operational improvement, deployment validation, incident management, and post-release review
Luxoft
DevOps & Cloud Engineering
- CI/CD pipeline development with deployment runbooks, rollback procedures, and production validation
- Release coordination across enterprise environments, infrastructure automation
DXC Philippines
Infrastructure & Automation
- Infrastructure delivery, cloud migration, and structured deployment planning across multiple clients
- Operational readiness, go-live support, and large-scale modernization
The portfolio projects on this site are how I make newer areas visible: by building, documenting, and testing them in public rather than claiming expertise I have not earned yet.
Current Focus
What I'm Currently Exploring
The direction I'm actively pushing further through case studies, prototypes, and production-inspired builds.
Edge computing
How smaller distributed deployments change operational design, observability, and failure handling when you cannot assume a single central environment.
Hybrid cloud
Patterns that split work between cloud services and self-hosted components without losing track of ownership, deployment flow, or support boundaries.
AI systems with data boundaries
Architectures where retrieval, inference, and storage have to respect clear rules about what data can stay local, what can move, and why.
Systems Thinking
Architecture Diagrams
Supporting diagrams for the systems above: topology, component layout, and data flow.
EKS Triage Copilot
Serverless AI triage on AWS
Homelab Infrastructure
Proxmox + K8s cluster
FrostStream CDC
PostgreSQL → Kafka → MinIO
OpenRAG Business
ROI & use cases
OpenRAG Technical
RAG platform architecture
Vector DB Explainer
RAG, embeddings & ANN deep-dive
FinOps V1 (Fargate)
ECS + ALB enterprise architecture
FinOps V2 (EC2)
Cost-optimized — 78% cheaper
About Me
Practical Engineering,
Clear Tradeoffs
I'm Gary Climacosa. My background is in enterprise infrastructure, automation, SaaS operations, and cloud delivery, with hands-on experience across AWS, Azure, and Kubernetes environments.
More recently, I've been using portfolio work to explore AI systems and hybrid architectures in a practical way: by deploying real components, documenting the tradeoffs, and keeping the infrastructure visible.
I don't present myself as a deep edge specialist yet. It's an area I'm actively exploring, especially where edge, hybrid cloud, and AI systems meet and where data boundaries shape the design.
What matters to me is building systems that can be explained clearly: where data lives, what is automated, what it costs to run, and how another engineer would support it later.
FinOps Foundation member since 2025
How I Work
Embedded with your team
I integrate directly with engineering teams — not as an outsider delivering a report, but as a contributor shipping alongside your people.
Pairs with developers & SREs
Infrastructure decisions are better when the people operating them are part of the conversation. I pair, review, and teach.
Simple, auditable, automated
Every system I build can be understood by the next engineer. No hidden complexity, no heroic runbooks.
Get In Touch
Let's Connect
Open to senior cloud, platform, and systems roles, especially where the work is hands-on and involves AWS, Azure, Kubernetes, hybrid cloud, or practical AI systems.
I aim to respond within one business day.