MTtargiel.proSystems & Integration Architect
Michał Targiel

Systems & Integration Architect

Salesforce·Enterprise Integrations·AI Systems

I design Salesforce platforms, enterprise integrations, and AI-enabled business systems — and build independent products around document processing and workflow automation.

View products & systemsArchitecture notesContact
~15 years in IT
engineering → architecture
Salesforce Architect
current role · Synairo
Katowice, PL
B2B · EN / PL
Michał Targiel
TRAJECTORY
├─ Software Engineer
├─ Salesforce Developer
├─ Technical Lead
└─ Salesforce Architect  ◀ now
   └─ systems · integrations · AI
01About

Engineer first.
Architect by craft.

I started in hands-on software engineering and spent roughly 15 years working across enterprise applications, ERP, .NET, SQL, integrations and Salesforce. Over time I moved through Salesforce development and technical leadership into architecture.

Today my work is increasingly focused on system design, integration boundaries, data flows and technical decisions across Salesforce and connected systems. In parallel, I build independent products around AI-assisted document processing and workflow automation.

I remain hands-on, but the direction is increasingly from implementing individual features toward designing systems, understanding their constraints and making explicit technical trade-offs.

02Current focus

Salesforce Architecture

├─ Solution design
├─ Salesforce platform architecture
├─ Data modelling
├─ Security & sharing
├─ Architecture reviews
└─ Asynchronous processing

Enterprise Integrations

├─ API design
├─ Salesforce integrations
├─ MuleSoft
├─ Event-driven workflows
├─ External systems
└─ Integration reliability

AI Systems

├─ Document intelligence
├─ Structured extraction
├─ Human-in-the-loop workflows
├─ AI evaluation
├─ AI + Salesforce integration
└─ Reliable production AI
03Products & systems
Own products
01 — built & owneddocai.synairo.com

DocAI

Verifiable AI Document Intelligence Platform

A document intelligence platform that transforms PDFs and images into structured business data using OCR, deterministic extraction, AI models, validation, and human review.

├─ Asynchronous processing with persistent jobs and retry handling
├─ Deterministic extraction first, probabilistic models second
├─ Human review as an explicit approval boundary
├─ Idempotent ingestion, signed webhooks, failed-event recovery
└─ Benchmarking of OCR and model configurations
AI SystemsArchitectureDocument IntelligenceAPI
View architecture case study →
Processing path
   Document
      │
   Ingestion            idempotent
      │
   OCR / Structured source
      │
   Extraction           deterministic → AI
      │
   Validation           confidenceHuman review         approval boundary
      │
   Approved data
      │
   Business system / Salesforce
02independent productscrum.synairo.com

Scrum

Workflow & Planning System

A planning and workflow system built end to end — problem definition, workflow design, application, and delivery.

Product EngineeringWorkflowInternal Systems
03built & ownedrzeczodruk.pl

Rzeczodruk

Independent product & business system

An operating 3D printing workshop with its own commerce and order system: product catalogue, categories, cart and accounts, a custom-project quoting flow, and fulfilment with stated lead times.

Product ownershipE-commerceOrder workflowOperations
Selected client work
Enterprise engagement · anonymized

Global Salesforce Platform

Enterprise Salesforce Architecture & Integrations

Solution design and technical architecture within a large global Salesforce environment, including integration design across systems of record and review of platform-level technical decisions.

SalesforceEnterpriseIntegrations
Scope of work
· Solution and integration design
· Data flows between Salesforce and external systems
· Platform constraints and limits
· Architecture reviews
· Maintainability and delivery across teams
BNB Academy of DanceProduction Web PlatformProduction web platform delivered for an operating dance academy, supporting class discovery, schedules, events, registration and studio information.Client DeliveryProduction
04Architecture notes

Decision records and notes written while building real systems. They describe the choices I made in a specific context, not universal rules.

reftitlestate
ADR 001Synchronous vs asynchronous document processingwritten ↓
ADR 002Why AI completion is not business approvaldrafting
Note 003Human review as a system boundarydrafting
Note 004Deterministic extraction before probabilistic AIdrafting
Note 005When image preprocessing improves OCR — and when it makes it worsedrafting
ADR 001DocAI · accepted

Synchronous vs asynchronous document processing

Context

Documents arrive through an API and a UI upload. A single document can require OCR, extraction, validation, and a human decision before its data is usable downstream.

Problem

A request-response call cannot hold open for the time this takes, and a failure mid-processing would lose work that has already been paid for in compute.

Constraints

External OCR and model calls have variable latency and can fail. Callers need immediate acknowledgement. Nothing may be processed twice as a side effect of a retry.

Options

Synchronous processing inside the request; asynchronous processing with an in-memory queue; asynchronous processing with persisted jobs and explicit state.

Decision

For this system I chose asynchronous processing with persisted jobs. Ingestion accepts the document, assigns an identity, and returns a job reference; every processing stage advances a stored state that survives a restart.

Trade-offs

More moving parts and more state to reason about, and callers must handle completion through polling or webhooks rather than a single response.

Consequences

Retries become safe because ingestion is idempotent on a client-supplied key. Failed stages can be resumed instead of restarted. Human review fits naturally as one more state rather than an exception path.

What I learned

Making the processing state explicit turned out to matter more than the choice of queue. Once each document had a durable state, reliability work became a matter of moving between states correctly.

received → queued → extracting → validating
        → review → approved → delivered
                     └─ failed → retry
05Experience

Software Engineer → Salesforce Developer → Technical Lead → Salesforce Architect. Each step moved further from single features and closer to how systems fit together.

periodrolework
2024 — presentSalesforce ArchitectSalesforce solution design and technical architecture in a global ~10,000-user environment, including integrations with SAP and legacy systems, data flows, platform constraints, and implementation guidance across Salesforce and connected systems.
2023 — 2024Technical LeadPlatform feature delivery and cross-system integrations; technical decisions and review for the delivery team.
2022 — 2023Technical LeadSalesforce CG Cloud delivery across retail and trade operations, working within platform constraints and data volumes.
2021 — 2022Technical Lead — Experience CloudGlobal Experience Cloud modernisation for 50,000+ portal users, including accessibility work (WCAG/ARIA).
2021 — 2022Salesforce DeveloperHR and partner community on Experience Cloud, integrated with SAP and legacy systems.
2019 — 2020Salesforce DeveloperMulti-channel CRM modernisation across several business units and brands.
2019 — 2020MuleSoft DeveloperIntegration engineering between Salesforce, SAP, and legacy systems.
2011 — 2019Software Developer (ERP / .NET / SQL)ERP systems, analytical tooling, and internal business applications — where the hands-on engineering habit comes from.
06Skills

Architecture

primary
Solution design·Integration architecture·Data modelling·Security & sharing·Async processing·Architecture reviews

Enterprise Platforms

Salesforce · MuleSoft · SAP integrations · Platform APIs

AI Systems

Document intelligence · OCR · Structured extraction · LLM integration · Human-in-the-loop · Evaluation

Engineering

Apex · LWC · Python · FastAPI · .NET / C# · SQL · REST · Git / CI/CD

CertificationsSalesforce Administrator·Platform App Builder·Platform Developer I·Agentforce Specialist·MuleSoft (DataWeave & Integration Basics)
07Contact

Open to conversations about Salesforce architecture, integrations, and AI systems.

Best fit: B2B engagements where Salesforce, integrations, or AI-enabled systems are part of the brief. Recruiters welcome.

Download CV
Send a message
Details
locationKatowice, Poland
languagesPolish (native) · English
engagementsB2B
response1–2 business days