FuseData is a data automation and engineering platform I help build at Noesys Software. The public product site at fusedata.cloud is in stealth while general availability is prepared; workflow capabilities are documented under Infoveave Automation v8.
What a data workflow is
A data workflow is a series of steps that define how data is collected, processed, transformed, and analysed. Automating those steps keeps pipelines organised, saves time, and improves accuracy — the same framing as Introducing Workflows in the help centre.
Workflow automation
Infoveave’s Automation module orchestrates connected Activities — extract, transform, validate, report generation, alerts — in a visual, no-code/low-code designer. Workflows can run on a schedule, on events, or manually, and integrate spreadsheets, databases, cloud storage, and third-party applications.
A sibling capability, Web Automation, simulates operational user actions in web apps (login, navigation, data entry, extraction) with validation and transformation against business rules. Workflow automation focuses on data-centric orchestration; web automation covers human-like UI interactions. Both sit on the same platform.
Activity ecosystem
Activities are the building blocks of automation (Introducing activities):
- ETL — SQL, file operations, database ops, transformations, REST/API calls
- Applications — connectors for Salesforce, Braintree, Dynamics, HubSpot, Shopify, Genesys, Mailgun, and more
- Data transformation — aggregation, formatting, geospatial transforms, filtering
- Email — Gmail, Outlook, send/move/download flows
- Flow control, RPA, Infoveave-native ops — packages, file transfer, Google services
What I build
At Noesys I work across the full lifecycle of data movement and transformation:
- Connectors — Salesforce, Braintree, Dynamics, GraphQL, OData; incremental ingestion, schema-aware extraction, resilient retries
- Transformation engine — mapping rules, schema normalization, JSON flattening, deduplication, validation
- Performance — DuckDB + Apache Arrow for columnar batch/micro-batch processing; warehouse integrations (BigQuery, Athena, Databricks, Snowflake)
- Orchestration UI — React + TypeScript workflow designer; CodeMirror, AG Grid, React Flow; ECharts for analytics views
- Platform core — .NET (gRPC/REST), SQLite for transient state, structured logging, metrics, CI/CD
- AI & extensibility — AI Agents and GenAI-driven automation paths; dynamic workflow execution (reflection/DI) so tasks can ship via assemblies without core changes; ML.NET pipelines for ML-ready preparation
- Observability — execution metadata, error routing, full data lineage; production safeguards on connectors
Resume highlights from this work: ~60% build-time reduction and ~90% project size reduction on deployment paths; ~40% memory footprint reduction using Apache Arrow on large workloads; led integration of DuckDB transformations into the workflow engine.