Skip to content
Filip Szczepanski

Independent consultant / Poland

Cloud infrastructure.DevOps. Automation.AI enablement.

Independent technology consulting for organizations modernizing infrastructure, cloud environments and technical workflows.

EDGEPLATFORMWORKLOAD
Reference topology — edge, platform and workload tiers
  • Azure
  • AWS
  • Google Cloud
  • DevOps
  • CI/CD
  • Infrastructure automation
  • AI enablement

Services

Five areas of work

Most engagements start in one of these areas and drift into the others as the picture gets clearer. They can be scoped as a short assessment, a defined project, or ongoing support.

SVC.CLD

Cloud & IT infrastructure

I help teams choose an architecture they can actually run: one that fits the workload, the budget, and the number of people who will maintain it after I leave. Usually that means Azure, AWS or Google Cloud, and often some combination of them.

  • Cloud infrastructure consulting
  • Infrastructure architecture and planning
  • Cloud environment assessment
  • Infrastructure modernization
  • Hybrid and cloud-native environments
  • Technical infrastructure strategy

SVC.MIG

Infrastructure migration

Migrations rarely fail on the main path. They fail on DNS, certificates, data volumes, and the one service nobody documented. I plan the route, run the cutover with your team, and stay through the weeks afterwards, which is when the real problems tend to show up.

  • Migration planning and execution
  • Hosting and cloud environment migrations
  • Migration architecture and sequencing
  • Cutover planning and technical coordination
  • Post-migration stabilization
  • Analysis and resolution of migration-related issues

SVC.DEV

DevOps & CI/CD

The goal is a release nobody has to plan their week around. I work on the pipeline and the process together, because a fast pipeline bolted onto a fragile process helps no one.

  • CI/CD architecture and optimization
  • Deployment pipeline design
  • Infrastructure deployment processes
  • DevOps consulting
  • Infrastructure automation
  • Improving deployment reliability and repeatability

SVC.AUT

Automation

Most teams already know which tasks eat their week. Fewer have had a spare afternoon to work out which of them are worth automating. I start with the ones that are risky to do by hand, not the ones that are easiest to script.

  • Infrastructure process automation
  • Deployment automation
  • Workflow optimization
  • Evaluation of automation opportunities
  • Reduction of repetitive technical processes

SVC.AI

AI enablement

AI is worth adopting in specific places and a distraction in others. I look at where it fits the work you already do, connect tools that are proven rather than promising, and hand back a process your team can run without me.

  • AI readiness assessment
  • Evaluation of AI tools and use cases
  • AI-supported workflows
  • AI-assisted technical and business processes
  • Integration of existing AI solutions into workflows
  • AI-supported verification and automation

AI enablement

What AI enablement actually involves

Most organizations do not need a model of their own. They need proven AI services connected to the systems that already hold their data, with the access rules, the running costs and the failure behaviour worked out before any of it reaches production.

AI.INT

Integration

Connecting AI services to the systems and data you already run, without opening holes in either.

  • Integration of existing AI services and APIs
  • Connecting models to internal data sources
  • Identity, access and permission boundaries
  • Cost controls, quotas and usage monitoring
  • Fallback behaviour when a model is slow or unavailable
  • Logging and auditability of AI-assisted steps

AI.IMP

Implementation & rollout

Getting from a demo that worked once to something people can rely on every day.

  • Pilot-to-production rollout
  • Evaluation criteria defined before launch
  • Human review and approval steps
  • Guardrails and output validation
  • Team enablement and internal documentation
  • Staged rollout and measurement of real usage

AI.SYS

System planning with AI

Designing systems where AI is one component among several, with explicit limits on what it decides.

  • Architecture for AI-supported systems
  • Deciding what AI should and should not handle
  • Data flow, retention and residency planning
  • Model and vendor selection, kept replaceable
  • Deterministic fallback paths
  • Latency, capacity and cost modelling

Where it earns its place

  • Summarizing and classifying large volumes of text
  • Drafting work that a person then reviews
  • Extracting structure from unstructured input
  • Search across internal documentation
  • Checks that are tedious but well defined

Where it usually does not

  • Decisions that need a reproducible, identical result
  • Processes with nobody available to review the output
  • Work a deterministic script already handles well
  • Cases where a wrong answer costs more than the time saved

Expertise

Platforms and areas

Platform recommendations follow the workload. Each of the major clouds fits some requirements better than others, and part of an estate is often best left where it already runs.

Cloud platforms

  • Microsoft Azure
  • Amazon Web Services (AWS)
  • Google Cloud Platform (GCP)
  • Hybrid environments

Infrastructure

  • Cloud infrastructure
  • Infrastructure architecture
  • Infrastructure as code
  • Environment assessment

Delivery

  • DevOps
  • CI/CD
  • Deployment pipelines
  • Infrastructure automation

Transformation

  • Infrastructure migration
  • Cloud migration
  • AI enablement
  • Technical process design

Regions worked in most often

  • eu-central-1
  • westeurope
  • europe-west3

Approach

How an engagement runs

Infrastructure work has an order to it. Skipping the early phases is what produces migrations that stall halfway and automation nobody trusts.

Engagement start

  1. Phase 1

    Assess

    Work out what is running today, what constrains it, and what the business actually needs it to do, including what it currently costs.

  2. Phase 2

    Design

    Define the architecture, migration strategy and technical processes. Decisions get written down so they can be reviewed by people who were not in the room.

  3. Phase 3

    Implement

    Hands-on work alongside your team: infrastructure changes, migration execution, automation and deployment processes.

  4. Phase 4

    Stabilize & optimize

    Resolve the issues that surface after a change lands, and improve reliability and efficiency once the system is under real load.

  5. Phase 5

    Enable

    Introduce automation and AI-supported workflows where they provide measurable value, and hand over processes the team can run without you.

Ongoing operation

About

Who you would be working with

I'm Filip Szczepanski. I work as an independent consultant on infrastructure, cloud environments, DevOps practice, and the practical side of adopting AI.

You work with me directly. There is no account manager in between, no rotating project team, and no handover between the person who scoped the work and the person who does it.

Most of what I take on starts with an IT department that has the context but not the spare capacity: a migration with a date already promised to the business, a deployment process that has outgrown the way it was first set up, or a question about whether AI is worth the budget before anyone commits to it.

I am based in Poland and work with organizations across Europe, usually from inside an existing team and its processes rather than alongside them.

You talk to the person doing the work
No layer in front of me, and no translation step between what you explain and what gets built.
One person, start to finish
I scope, design and implement, so nothing gets lost at a handover that never happens.
Decisions get written down
Architecture choices come with the reasoning attached, so it is still there once I have gone.
The boring option, usually
The simplest thing that meets the requirement. Interesting is not a goal.
No platform loyalty
I recommend platforms based on your workload and your team, not on a partnership agreement.

Contact

Planning a cloud migration, infrastructure transformation or AI enablement initiative?

Tell me what you are running now, what needs to change, and any dates that are already fixed. That is usually enough for me to say whether I can help and how I would go about it.

Direct

contact@szczepanski.us

Email works just as well. I read and answer these myself.

Project enquiry

Your details are used only to answer this enquiry.Privacy policy