07/23/2026 updated
100 % available
Senior Salesforce Data Cloud & MarTech Engineer | Marketing Cloud, ETL, API
Valencia, Spain
Only remote
BE, Computer ProgrammingAbout me
Salesforce Data Cloud, SFMC and CRM. I turn fragmented customer data into unified, activation-ready models — ingestion, harmonization, identity resolution — and back it with solid data engineering (ETL, SQL, Python, APIs). Valencia-based, 100% remote, EU work authorization.
OAuth 2.0RESTful API Modeling LanguageExtract Transform Load (ETL)Data ModelingData VisualizationPython (Programming Language)Salesforce Marketing CloudSAS (Software)Simple Object Access Protocol (SOAP)WorkflowsApex CodeSQL OptimizationSalesforce LightningAdvanced ReportsSalesforce Object Query Language (SOQL)
Senior Data & MarTech Engineer — Salesforce Data Cloud and Marketing Cloud architect, developer and QA on a full-stack banking implementation.
SALESFORCE ARCHITECTURE & DELIVERY
- Data Cloud (Data 360): 50+ data streams, DMO modeling, identity resolution, calculated insights, segmentation and activation to SFMC — from data-dictionary design to production. Built an Apex stream-health monitor on the Data Cloud Query API v2 that detects stale streams, coverage gaps and data-quality anomalies across the whole ingestion layer.
- Solved priority-offer selection that Calculated Insights cannot express: a zero-padded composite ranking key as a Data Stream formula field, returning one top offer per client in activation (~395k audience).
- SFMC: journeys, automations, Email Studio, AMPscript, SQL activities, Contact Builder. Designed a cross-journey control-group architecture (master DE + protection window) after auditing 57k clients contacted inside another campaign's control period.
- Platform governance: contact-license and Super Message consumption audit (665k contact cap, 6M sends / 6 months), deletion strategy, deliverability and DKIM setup.
- Salesforce CRM: Apex, SOQL, FSC data model, Flows, FLS / Profiles / Permission Set Groups, Lightning pages — hands-on admin, developer and QA on a live banking org.
INTEGRATION & CUSTOM DEVELOPMENT
- Custom SFMC Journey Builder activities for SMS, Viber/SMS, Push and Call Center over a Java/Spring Boot backend (Kong, PostgreSQL, Kubernetes): configuration UI, personalization validation, async delivery callbacks, sync into Data Extensions and Data Cloud.
- Defined the cross-channel delivery-status taxonomy from 90 days of production data (terminal vs non-terminal statuses, Viber→SMS cascade), turning inconsistent callbacks into a reliable reporting funnel.
- REST/SOAP integrations, OAuth2, MuleSoft-fed streams, SFMC↔CRM consent write-back.
ANALYTICS & DOCUMENTATION
- SQL (SFMC T-SQL dialect, Data Cloud SQL), Python; funnel and engagement analysis on multi-million-row datasets.
- Einstein / Data Cloud ML PoC design: propensity models, feature preparation, exclusion of privacy-proxy open noise from training targets.
- Author of the client-facing design and analytics documentation (Confluence, English) the bank's teams work from.
WORKING MODE
Architect, developer and QA in one role on a full-stack banking implementation. Business-to-technical translation, direct client communication. Valencia, Spain — 100% remote, EU work authorization. English B2.
Experience
Professional experience5 - 10 years
Freelancer since2025
Languages
BulgarianGoodEnglishGoodRussianNative speaker
Project history
Retail bank, full-stack Salesforce: Data Cloud, Marketing Cloud, FSC and acustom Java notification backend. I was the architect, the developer and the QAon it.
The work that mattered was the work the platform refuses to do. CalculatedInsights cannot rank offers — I moved the ranking into a composite sort key onthe data stream and shipped it to a 395k activation. Data Cloud has no streammonitoring — I wrote one in Apex on the Query API, and stale streams stoppedsurfacing as broken campaigns. Control groups were honoured inside a journey butnot across journeys — an audit found 57k clients contacted during anothercampaign's protection window, and I designed the governance that closed it. SFMChas no Viber or Push — I built the Journey Builder custom activities and thedelivery-status model behind them, from 90 days of real callbacks.
I then defined and costed the bank's next phase: Einstein propensity models onData Cloud (feature preparation, target design, lift as the success metric), apaid-media layer via Marketing Cloud Advertising — retargeting, control-groupsuppression, lookalike — with the retirement path to Data Cloud Ad Audiencesbuilt in from day one, and a template-manager app for the custom channels. Eachdelivered as a design the bank can approve and a number it can budget against.
Tech: Data Cloud, SFMC, AMPscript, Apex, SOQL, FSC, SQL, Python, REST/SOAP,
The work that mattered was the work the platform refuses to do. CalculatedInsights cannot rank offers — I moved the ranking into a composite sort key onthe data stream and shipped it to a 395k activation. Data Cloud has no streammonitoring — I wrote one in Apex on the Query API, and stale streams stoppedsurfacing as broken campaigns. Control groups were honoured inside a journey butnot across journeys — an audit found 57k clients contacted during anothercampaign's protection window, and I designed the governance that closed it. SFMChas no Viber or Push — I built the Journey Builder custom activities and thedelivery-status model behind them, from 90 days of real callbacks.
I then defined and costed the bank's next phase: Einstein propensity models onData Cloud (feature preparation, target design, lift as the success metric), apaid-media layer via Marketing Cloud Advertising — retargeting, control-groupsuppression, lookalike — with the retirement path to Data Cloud Ad Audiencesbuilt in from day one, and a template-manager app for the custom channels. Eachdelivered as a design the bank can approve and a number it can budget against.
Tech: Data Cloud, SFMC, AMPscript, Apex, SOQL, FSC, SQL, Python, REST/SOAP,
OAuth2, MuleSoft, Einstein.
Migrated a bank's marketing off SAS Real-Time Decision Manager onto CM Ocean, a real-time CVM platform, in six months: campaigns, customer segments and the data underneath them, Oracle to Postgres and Tarantool, with five complex decisioning flows re-implemented rather than lifted.
The platform covered the standard cases. Everything else I wrote myself — business logic as custom Python modules, deployable and callable directly from a campaign, so marketers could run rules the product had no screen for. I configured the backend, built the images and ran the Kubernetes deployments myself instead of queueing behind a platform team.
Campaign triggers arrived over Kafka with no visibility into whether they landed. I built topic monitoring that tracked input-parameter consumption, caught failures early and restarted campaigns automatically through a custom API module — silent trigger loss became a self-healing path.
Replaced manual functional and load testing with automated frameworks, and built the onboarding infrastructure the team ran on, so new engineers stopped depending on whoever happened to know.
Tech: Python, Kafka, Airflow, Docker, Kubernetes, SQL, Oracle, Postgres, Tarantool, ETL, pandas, scikit-learn, A/B testing.
Delivered real-time marketing decisioning for a top-tier retail bank — tens of millions of customers, heavily regulated environment — on SAS Real-Time Decision Manager and SAS Marketing Automation.
Took 15+ decisioning cases from campaign brief to production and owned them once they were live: offer recalculation, card issuance, instalment products, transactional triggers. Each meant reading the business intent, turning it into decisioning logic against live customer data, and defending that logic when real traffic hit it.
Testing was the part nobody had solved. Verification meant reading logs by hand, one scenario at a time — which is why defects surfaced in production rather than before it. I automated it with SoapUI and Postman, running scenarios in parallel with structured, comparable output, then extended the same harness to load simulation so we knew how the system behaved at campaign-peak traffic before a campaign created it.
This is where the habit started that I have kept since: build the instrumentation around a platform, not just the configuration inside it.
Tech: SAS RTDM, SAS Marketing Automation, SQL, SoapUI, Postman, Python, load testing, A/B testing.