★ 4
Sparquet
Open-source data engineering framework for Apache Spark. Every pipeline is one declarative JSON contract: write it, generate it with any LLM, or design it visually in Sparquet Studio.
Apache SparkPythonDelta Lake
Data Engineer
Victor
Pasqualini
Data Engineer building large-scale pipelines in the cloud. I migrate legacy systems, process terabytes a day with Spark, Kafka, Airflow and Databricks, and deliver consistent pipelines on AWS and GCP, with integrity and quality end to end.
Data Specialist.
I have been a Data Engineer for more than five years, designing and running pipelines in cloud architectures. I work on moving legacy systems, from on-premise Cloudera, Hadoop/Hive estates to Cobol/IBM DB2 mainframes, onto modern data platforms such as Databricks, with ETL/ELT processing in Spark and infrastructure on AWS and GCP.
Across credit bureaus, capital markets and consultancies I have built medallion lakehouses over 20+ TB of daily ingestion, cut batch processing time by 40% on Databricks, and held a 90% SLA on pipelines other teams depend on. Data quality, integrity and availability are the part of the job I care about most, and open source is what I build it with.
Core stack
SparkKafkaAirflowDatabricksIcebergDelta LakePythonScalaSQLNoSQLAWSGCPTerraform
Certifications
Confluent
MongoDB
Experience
Led a complex mainframe (Cobol/IBM DB2) to AWS pipeline migration processing massive daily positional file ingestion. Implemented data validation and quality with consolidated table generation (bronze/silver/gold) for ML and other products. Production deployment with dual-running across environments, integrations via EventBridge, REST APIs and IaC with CloudFormation.
Spark (Scala/Python)DuckDBPandasPolarsPySparkEMREC2LambdaSQSMSKMWAAS3IcebergDelta LakeAthenaRDS/AuroraDynamoDBDocumentDBGlue CatalogDatadogGrafanaCloudWatch
What I've built
★ 3
Sparquet Cola
Data quality for Spark that runs where your data already is. SODA-style metric checks, SQL rules, schema contracts, plus a valid/invalid split that tells you which rule rejected each row. Pure PySpark, no extra services.
Apache SparkPythonSQL
Pulse
A personal finance dashboard that reads the spreadsheet you already keep on OneDrive. Income and expenses by month, spending by segment and by week, credit card, and a screen just for investments, where contributions don't count as spending and yields don't count as income. No database, no new format: it maps the columns you already use.
Next.jsReactTypeScript
Pouch
A multi-strategy crypto bot built around the premise that almost nothing works: it sweeps strategies, validates every survivor on history the optimiser never saw, and only then trades them on the Binance Spot Testnet (real API, fake money).
PythonLightGBMFastAPI
The day-to-day stack
Languages
Python (PySpark, Pandas, Polars, DuckDB)
Scala
JavaSQL
Cobol
RustETL / ELT
Spark (Batch/Streaming)
Flink
Glue
EMR
EC2 (spot)
Lambda
Sqoop
DataFusion Comet
DatabricksPhoton
Snowflake
dbt
Compute Engine
Cloud FunctionsIncrementalCDCMessaging
Kafka
SQS
MSK
RabbitMQ
RedpandaOrchestration
Airflow
MWAA
EventBridge
Step Functions
Composer
Ctrl-M
NiFiDatabricks DLTDatabricks WorkflowsStorage & Lakehouse
S3Glue Catalog
Iceberg
Hadoop
Kudu
BigQuery
Cloud StorageDelta LakeUnity CatalogMedallion ArchitectureDatabases
RDS/Aurora
PostgreSQL
MySQL
DynamoDBMongoDB
DocumentDB
OracleIBM DB2Infra
Terraform
Terragrunt
CloudFormationAnalytics & Observability
Datadog
Grafana
Metabase
Power BI
HiveImpala
ElasticSearch
CloudWatch
AthenaOther
CI/CD (Cockpit)
Git (GitLab, Bitbucket)
Linux (Shell/Bash Script)
Skill Badges
Degrees
2025 — 2026
Inbix Academy
MBA - AI for Innovation
2024 — 2025
FIAP
MBA - Data Engineering
2020 — 2022
FATEC São Paulo
Bachelor's, Systems Analysis and Development
2017 — 2019
ETEC Jardim Ângela
Technical, Computer Science (integrated high school)
Get in touch
Let's talk dataover a coffee? ☕
A pipeline that won't close, a migration stuck halfway, a role you're hiring for, or just the urge to talk data. Reach me on whichever channel you like.
São Paulo, Brazil--:--Local time