Technologies Tested in Real Production
As a Data Platform Architect and Senior DBA with 29 years of hands-on experience, I choose tools based on fault tolerance, predictable operational costs, and raw performance under load. Below is an unvarnished assessment of the database engines, multi-cloud platforms, automation frameworks, and AI systems powering my production environments.
Oracle Database
Mission-critical enterprise transactional environments, high-throughput OLTP processing, high availability with Oracle RAC, and disaster recovery via Data Guard.
Financial institutions, insurance corporations, telecommunications providers, and large enterprises demanding uncompromising reliability and full ACID compliance.
High per-core licensing expenditure; intricate memory architecture (SGA/PGA) and query tuning demanding decades of hands-on engineering mastery.
Enterprise commercial licensing (per processor core factor + annual support)
Microsoft SQL Server, PostgreSQL, IBM Db2
The primary pillar of my 29 years of experience. I served as L3 engineering support and architect for hundreds of mission-critical Oracle databases across global enterprises (including Kyndryl / IBM, Aviva).
Microsoft SQL Server
High-performance enterprise relational database with battle-tested High Availability (Always On Availability Groups) and robust Disaster Recovery architecture.
Organizations built on the Microsoft ecosystem, enterprise data warehouses, mission-critical ERP/CRM transactional platforms, and corporations with strict SLAs.
Enterprise Edition demands substantial licensing investment; multi-subnet Always On clustering requires meticulous network tuning, quorum configuration, and listener routing.
Commercial licensing (per-core model for Standard / Enterprise editions)
Oracle Database, PostgreSQL, Azure SQL Managed Instance
Managed an estate of over 2,000 MS SQL Server instances across global environments, overseeing Always On clusters, slashing RTO/RPO metrics, and resolving complex storage I/O bottlenecks.
PostgreSQL & pgvector
Modern open-source relational database featuring native vector storage for semantic search (RAG), declarative partitioning, and high availability clustering with Patroni.
Data architects, teams engineering sophisticated web platforms, and AI engineering projects seeking reliable long-term vector memory without siloed niche databases.
Demands rigorous tuning of autovacuum and connection pooling (PgBouncer) under heavy concurrency; HNSW vector indexes require substantial RAM allocations.
Open-source (free) / Managed Cloud (RDS, Cloud SQL, Neon, Supabase)
Oracle Database, Microsoft SQL Server, MySQL
Core database foundation at DBAdmin, persistent long-term memory for autonomous AI agents, and my primary target for cost-saving migrations away from proprietary commercial engines (e.g. Oracle to PostgreSQL).
MySQL
Proven and dependable relational engine (InnoDB) optimized for read-heavy workloads, read-replica scaling, and high-traffic web applications.
Web platforms, e-commerce stores, SaaS architectures, and cloud environments requiring a battle-hardened relational engine with an accessible operational curve.
Less expressive stored procedures and analytical window functions than PostgreSQL or Oracle; asynchronous replication can incur replication lag under heavy write bursts.
Open-source (GPL) / Commercial Enterprise (Oracle)
MariaDB, PostgreSQL, Amazon Aurora MySQL
A quintessential engine in internet architecture. Consistently maintained in my active production portfolio and utilized in workloads requiring rapid, low-latency query throughput.
MariaDB
Fully open-source relational database featuring synchronous multi-master clustering via Galera Cluster, columnar analytics with ColumnStore, and Oracle PL/SQL compatibility mode.
Engineering teams seeking an independent, highly available SQL engine with zero vendor lock-in and a frictionless modernization path away from Oracle.
Divergent syntax features and architectural changes relative to newer MySQL 8.x releases can complicate bidirectional migrations.
Open-source (GPL v2) / MariaDB Enterprise
MySQL, PostgreSQL, Percona Server
An outstanding choice for environments requiring multi-node synchronous clustering without commercial licensing overhead, offering SQL_MODE=ORACLE to run legacy PL/SQL packages natively.
MongoDB
Document-oriented NoSQL database (BSON) designed for semi-structured data, agile polymorphic schemas, and horizontal sharding across distributed clusters.
Telemetry ingestion pipelines, dynamic e-commerce product catalogs, mobile backends, and distributed systems with rapidly evolving document models.
Demands disciplined schema design (embedding vs. referencing trade-offs); cannot fully substitute classic relational engines for deep multi-entity analytical joins.
Community Edition (SSPL) / Managed MongoDB Atlas
PostgreSQL (JSONB), Amazon DocumentDB, Couchbase
I design and maintain MongoDB clusters as complementary layers to relational architectures wherever document structures fundamentally simplify data access and ingestion pipelines.
Amazon Web Services (AWS)
Comprehensive enterprise cloud infrastructure: managed databases (RDS Multi-AZ, Aurora), scalable compute (EC2), durable object storage (S3), and zero-trust VPC networking.
Enterprises requiring resilient, globally distributed cloud infrastructure with automated operations and high-availability guarantees.
Complex data egress billing and layered IAM permission models; necessitates continuous FinOps cost auditing and Savings Plans / Reserved Instance optimization.
Pay-as-you-go / Savings Plans / Reserved Instances
Google Cloud Platform, Microsoft Azure, IBM Cloud
A foundational platform in my enterprise automation initiatives. At Dun & Bradstreet, I automated multi-cloud operations across AWS and GCP, reducing repetitive DBA toil by 40%.
Microsoft Azure
Enterprise cloud platform featuring seamless native integration with Microsoft SQL Server, Azure SQL Managed Instance, Entra ID (Active Directory), and hybrid datacenters.
Enterprises modernizing traditional on-premises database clusters into the cloud, and organizations deeply invested in the Microsoft enterprise ecosystem.
Layered licensing models (Azure Hybrid Benefit); requires precise capacity planning across vCore compute tiers and Premium SSD storage IOPS profiles.
Pay-as-you-go / Azure Reserved VM Instances
AWS, Google Cloud Platform, IBM Cloud
The linchpin for secure cloud migrations of MS SQL estates, ensuring uninterrupted high availability (Always On) and unified enterprise identity governance.
Google Cloud Platform (GCP)
Next-generation cloud infrastructure powered by Google's global private network backbone, managed database engines (Cloud SQL, AlloyDB), and scalable Big Data / AI platforms.
Modern data platforms and organizations building high-scale analytical data warehouses, machine learning workflows, and multi-cloud resilience.
Distinct organization and project resource hierarchy compared to AWS; fewer specialized enterprise legacy tooling options for older SQL Server editions.
Pay-as-you-go / Committed Use Discounts (CUD)
AWS, Microsoft Azure
Implemented at Dun & Bradstreet alongside AWS within a robust multi-cloud architecture, delivering agile automated instance provisioning and data workloads.
IBM Cloud
High-security enterprise cloud engineered for rigorous financial regulatory compliance, data sovereignty, and mission-critical hybrid workloads.
Financial institutions, global banking giants, insurance leaders, and regulated public sector organizations demanding stringent compliance certifications.
Smaller catalog of commodity developer services compared to hyperscalers (AWS/GCP); higher contract and enterprise procurement threshold.
Enterprise contracts / Cloud catalog pay-as-you-go
AWS, Microsoft Azure, On-Premises Datacenter
An environment directly tied to my tenure at Kyndryl / IBM, where I managed core database architecture and resiliency for premier global financial accounts.
Python & Ansible (DBA Automation)
End-to-end database infrastructure automation at scale: automated instance provisioning, recurring fleet patching, security parameter auditing, backup verification, and health monitoring.
Database administration teams and platform engineers aiming to eliminate repetitive manual toil and eliminate human error in operational workflows.
Demands strict idempotency and rigorous pre-production playbook testing prior to executing changes against live production clusters.
Open-source (free) / Optional Red Hat Ansible Automation Platform
Terraform, Bash / PowerShell, Puppet
The automation backbone I engineered at Dun & Bradstreet, delivering a 40% reduction in manual DBA tasks. I use Python extensively for deep performance telemetry diagnostics and query execution plan parsing.
TinaCMS
Git-backed headless CMS delivering real-time visual editing, structured content modeling, and direct commits to GitHub.
Developers and content creators who demand git version control, transparent pull requests, and absolute freedom from proprietary CMS lock-in.
Requires SSR deployment configuration or preview proxies for real-time visual editing in production environments.
Open-source for self-hosting / TinaCloud free tier & commercial tiers
Decap CMS, Sanity, Strapi
Provides the ultimate synergy: developers keep pure Markdown/MDX in Git repositories, while non-technical editors enjoy intuitive visual forms and instant previews.
AI Agents (OpenClaw & Paperclip AI)
Autonomous AI agent architecture executing complex analytical workflows: error log parsing, database schema inspection, and operational workflow automation.
Platform architects and systems engineers looking for verified AI intelligence in incident root-cause analysis, documentation synthesis, and deterministic automation.
Requires secure sandboxed execution runtimes and zero direct access to sensitive customer production data (strict Human-in-the-Loop gates for DDL/DML changes).
Model LLM API consumption (pay-as-you-go token pricing)
Claude Code, CrewAI, LangGraph
Deepening applied AI architectures and autonomous agent orchestration via OpenClaw and Paperclip AI, building hardened, production-grade bridges to enterprise data platforms.
Claude Code & MCP
Autonomous agentic coding environment equipped with secure, granular tool access via the Model Context Protocol (MCP).
Software architects and platform engineers building advanced automation systems, protocol bridges, and multi-agent development workflows.
Demands precise context window management and strict least-privilege tool execution permissions.
Anthropic API usage (pay-as-you-go token billing)
OpenAI Assistants, Cursor, Aider
Primary cognitive engine leveraged for CAPCOM deterministic state orchestration at LabAI.
Astro
Ultra-fast, content-driven web architectures with zero client-side JavaScript by default, optimized for SEO, Core Web Vitals, and readability.
Editorial engineering teams, expert knowledge hubs, and high-performance corporate web platforms.
Less optimal for heavily reactive, state-dense SaaS client applications than traditional SPA frameworks.
Open-source (free, MIT license)
Next.js, Remix, Nuxt
Core web framework powering the entire paweldomanski.eu ecosystem and our 4 brand sites.