Enterprise / On the field
When high traffic, teams and compliance are in play, heavy tools prove their worth. At scale, complexity isn't a cost — it's an investment.
KubernetesMulti-cloudTerraformVaultPrometheus + Grafana
Uses · Working System
Not the tools I use, the system I work with.
This page is not just a "which app do I use?" list. It is a summary of the working order I have built across system architecture, network, server infrastructure, software and operations. For me, picking a tool is part of the speed, trust, scalability and sustainability decision.
MacBook Pro M4 Pro
Terminal-centric
Multi-cloud
AI-assisted
Working Context
Two scales, two disciplines
Tool choice depends on context. At enterprise scale, heavy machinery earns its place; in my own projects, radical leanness is an engineering discipline — this blog included.
In my own projects
Here the rule flips: the fewest moving parts win. This blog runs on a single VPS — Astro + Node + SQLite + Nginx + systemd, atomic deploys via a GitHub Actions self-hosted runner, a multi-AI content pipeline, and Umami analytics (no Google). "One VPS is enough" isn't a slogan — it's a working proof.
Astro + NodeSQLiteNginxsystemdGitHub Actionsmulti-AIClaude CodeUmami
Speed First
Every tool I use must shorten the time between decision and execution. Unnecessary clicks, slow startups and cluttered interfaces are negative points for me.
Operational Reality
For me, the value of a tool is measured by how much trust it earns under real load. I do not prefer flashy but fragile setups.
Software + Infrastructure Together
Editor, terminal, cloud and security tools are not separate worlds. All of them are picked as parts of the same working system.
Hardware
Main Machine
MacBook Pro Mac16,7 (2024)
The center of my daily workflow. Writing code, terminal-based operations, infrastructure management, documentation, content production and AI-assisted development — everything flows through the same machine. A unit where I almost never hear the fan and that stays thermally stable.
- CPU: Apple M4 Pro · 14-core CPU (10P + 4E)
- GPU: 20-core · Hardware-accelerated ray tracing
- Memory: 24 GB unified memory · 273 GB/s bandwidth
- Storage: 1 TB SSD · ~7 GB/s sequential read
- Display: Liquid Retina XDR · 3024×1964 · 120 Hz ProMotion
- External Monitor: Lenovo Legion R27fc-30 · 1920×1080 · 165 Hz
- Network: Wi-Fi 6E · Bluetooth 5.3 · 10 Gbit/s Ethernet (Thunderbolt)
- OS: macOS 15 Sequoia · Homebrew + mise
Desk Setup
There is nothing unnecessary on my desk. Each piece is picked to reduce the operational load of long hours. Meetings, deep-focus work and on-call interventions all flow through the same setup.
-
Keyboard
Magic Keyboard + Keychron K8 Pro (mechanical backup)
-
Mouse
Logitech MX Master 3S · fast switching between 2 devices
-
Audio
Sony WH-1000XM5 · noise cancelling + long meetings
-
Microphone
Shure MV7+ · podcast/meeting recording quality
-
Camera
MacBook built-in + Continuity Camera (iPhone)
-
Light
Elgato Key Light Air · neutral white 4500K
-
Stand
Monitor mount + adjustable laptop stand
-
UPS
APC Back-UPS · 15 min against power outages
Software Stack
Layers of My Daily Work
Development Console
The layer where I write code and interact directly with the systems. Speed, focus and low friction are decisive for me here.
VS Code / Cursor
Daily development surface
My first choice for fast navigation inside a codebase, AI-assisted generation and lightweight workflows. Main extensions: GitLens, Error Lens, ESLint, Astro.
JetBrains IDEs
Deep project work
I use these on the Go, Python and Java side for strong analysis and refactoring on large projects. GoLand + PyCharm is the core duo.
iTerm2 + Zsh
Terminal headquarters
The main backbone of SSH, log analysis, deployment, automation, service control and daily operations. Oh My Zsh + powerlevel10k prompt.
Git + GitHub
Version and flow discipline
Indispensable for keeping code quality, change history and team coordination under control. Commit signing + conventional commits.
Infrastructure and Cloud Layer
Since I build large structures, the tools I use are picked not just for deploys, but for scale, security, observability and continuity.
Cloudflare
Edge, DNS and cache layer
DNS, CDN, cache and WAF security layer. This blog runs on my own VPS (Astro + Node); Cloudflare sits in front purely as edge cache + protection — not Pages/Workers. On the enterprise side, Workers/D1 are in play too.
AWS / GCP / Azure
Multi-cloud workspace
Depending on the job, this gives me flexibility on services, cost, availability and integration. I do not stay tied to a single cloud.
Kubernetes + Docker
Run-time standard
I use these to make container-based services scalable and manageable. Helm + Kustomize packaging.
Terraform
Managing infrastructure as code
It supports the repeatable, auditable and portable infrastructure approach instead of manual setup.
GitHub Actions
CI/CD backbone
I use it to standardize build, validation, release and automation jobs.
Nginx
Reverse proxy and web layer
On my own servers I prefer it for SSL termination, load balancing and serving static content. This blog also runs behind it.
Software and Automation Languages
When picking a tool, the language is not what matters — the problem is. Still, some languages naturally sit closer to the center of my work.
Go
Services and CLI work
Very strong for backend and tool development thanks to performance, easy distribution and clean binary output.
Python
Automation and productivity
Very practical for scripting, data processing, integration and AI-adjacent tasks. uv + ruff is the modern toolchain.
TypeScript
UI and modern web layer
Provides a safe development flow in the frontend and Node ecosystem. Strict mode by default.
Astro
Content-focused publishing system
Offers a clean and fast foundation for performance and content-centric setups like this blog. Content Collections + MDX.
Observability and Security
I cannot manage a system without seeing its state; I cannot protect what I cannot see. This layer is a lifeline for me.
Prometheus + Grafana
Metrics and visualization
For years, my main metrics stack across both physical and cloud infrastructures. Exporters + dashboard discipline.
Loki / Elasticsearch
Log aggregation
Loki is light at scale, Elasticsearch is for deep search. Log structure is designed up front.
eBPF + Cilium
Kernel-level observation
A modern kernel toolset for network flows, syscall tracing and security policy enforcement.
Vault / 1Password
Secret management
Application secrets in Vault, personal and shared accounts in 1Password. Never plain text, ever.
AI Working Layer
I use AI tools not for show, but to increase my thinking speed and production quality. The final call still belongs to architectural judgment and field experience.
Claude Code
Production power inside the terminal
A very strong assistant for deep work inside a codebase, editing, reasoning and fast iteration.
Claude
Analysis and technical thinking support
I use it in architectural discussions, system design, content production and review flows.
Gemini API
Automatic content generation
The hourly publishing pipeline of this blog is based on Gemini models. Disciplined model fallback + retry.
GitHub Copilot
Micro speed boost
Useful in repetitive code, boilerplate work and keeping my flow uninterrupted.
Productivity and Knowledge
Not just software and infrastructure; the layer of organizing thought and fast access is also part of the same system.
Raycast
macOS launcher and quick commands
It replaced Spotlight. Snippets, clipboard history, window management and AI integration.
Notion
Notes and documentation
Architectural decisions, project notes and knowledge base. Practical for shared documents too.
Obsidian
Personal knowledge graph
Local markdown-based. A linked-thought system that connects my own ideas to each other.
Arc / Safari
Browser
Safari for everyday browsing (efficient + secure), Arc workspaces for separating development and work.
Terminal
Shell & CLI Lineup
Shell Setup
Shell
Zsh 5.9 + Oh My Zsh
Prompt
powerlevel10k
Active Plugins
gitdockerkubectlterraformawsgcloudfzfzsh-autosuggestionszsh-syntax-highlighting
Killer CLI Tools
- fzf
Fuzzy finder — files, command history, git branches
- ripgrep (rg)
A faster grep — sub-second even across millions of files
- bat
cat + syntax highlighting + paging
- eza
ls replacement — git status and icon support
- zoxide
Smart cd — jump to frequent directories with one command
- jq / yq
Swiss army knife for JSON and YAML transformations
- htop / btop
System monitoring — modern incarnations of top
- mtr
Network troubleshooting — traceroute + ping combined
Home Network
Home Lab & Network
As a network specialist, I want the setup at my own home to reflect the same discipline. The lab I run at home is a small model of the VLAN, segmentation and observability practices I apply in the field.
-
Router
pfSense · VLAN segmentation + DNS filtering
-
Switch
MikroTik CRS326 · managed L3 + PoE
-
AP
UniFi 6 Pro · Wi-Fi 6 + mesh
-
VPN
WireGuard · remote access + site-to-site
-
NAS
Synology · ZFS-like snapshot discipline
-
Monitoring
Prometheus + Grafana · home service health
Speed
Shortcuts I Use Often
⌘ ⌥ Space Open Raycast
⌘ K Blog site-wide search
⌘ T New terminal tab
⌃ R Shell command history (fzf)
⌘ P VS Code file navigator
⌘ ⇧ F Global search across the codebase
⌥ ←/→ Switch Arc workspace
⌘ ⌃ Space Emoji / character picker
Subscriptions
What I Pay For Out of Pocket
I also use free options, but I find these services worth paying for because they directly add to productivity or quality.
Claude Pro
Writing, design, reasoning
GitHub Copilot
In-editor autocompletion
JetBrains Pro
All IDE rights
Cloudflare Pro
Edge + security + WAF
1Password Family
Secret management
Raycast Pro
AI, cloud sync
Notion
Documentation and notes
Domain + VPS
mehmetsari.ai + .com.tr · my own server
Philosophy
A good tool, for me, is one that reduces the workload
I prefer simple, fast, reliable and reusable setups over flashy but fragile tools. Especially on the system and network side, the real value of a tool is measured by how much visibility and control it gives you in a problem situation.
On the software side, I build flows that increase production speed without lowering quality. AI is an accelerator here, not a decision maker.
Even this blog is a live model of my own thesis: hourly automatic publishing, atomic deploy, hosting on my own server, daily backup. I have arranged my tools to reflect my own operational discipline.
Workflow
Same 4 steps in every project
STEP 1
Think
I do not pick a tool before understanding the need. First, the scale, risk, dependencies and operational load are clarified.
STEP 2
Build
Infrastructure, code and release flow do not move forward in isolation. They are designed together from the beginning.
STEP 3
Automate
Every manual repetition produces errors over time. I move whatever I can into automation.
STEP 4
Observe
A system that is built but not observed is incomplete. Health, log and behavior visibility is mandatory.
Note
This page is not static — it is a living operations dashboard
As new projects, new needs and new ways of working come up, the tools I use also change. But what does not change is my approach: speed, security, sustainability and choosing based on real field needs.
Last updated: June 2026