Better Together
PlexiFact
+ Databricks
Databricks is a powerful lakehouse + compute platform. PlexiFact is the fund-data layer. Many of our customers run both - PlexiFact reads and writes Delta Lake, and can delegate compute to Databricks SQL Warehouses or Photon while keeping fund-data orchestration on top. This page is about where each shines, not which one "wins."
- Domain-native NAV, lineage, multi-currency, recon - sits cleanly on top of the lakehouse
- Bidirectional Delta Lake - read Delta tables in, push curated outputs back
- Unity Catalog + Photon friendly - lineage hands off, compute delegates
- 01 LAKEHOUSE Delta Lake · Unity Catalog
- 02 COMPUTE Photon · Spark · SQL Warehouse
- 03 ML / AI MLflow · model serving · feature store
- 04 NOTEBOOKS Data science workspaces
- Compute + ML + lakehouse storage
- 01 INGEST 24+ domain connectors (Bloomberg, Citco, BNY)
- 02 RECONCILE Multi-prime breaks · positions · NAV
- 03 GOVERN Lineage · audit · domain-aware controls
- 04 DELIVER LP reporting · portfolio aggregation · APIs
- Fund-data layer on top of (or alongside) the lakehouse
Where Each Shines
Side by Side
Different layers of the stack, different strengths. Most of our customers use both.
Sweet Spots
Where Each Shines
Neither tool is "better" - they live in different layers. Here is how we think about scope.
Databricks is the right answer when
Lakehouse + ML + Spark compute
- Heavy data science and ML workloads (factor models, alpha research, NLP on filings/news)
- Petabyte-scale analytics workloads with bursty compute patterns
- Teams with dedicated platform engineering who model the lakehouse layer in-house
- Organizations standardizing on a single open-format compute platform across multiple business units
PlexiFact is the right answer when
Fund-data layer + connectors
- Alt asset managers needing a fund-data layer fast (LP reporting, multi-prime recon, portfolio aggregation)
- Teams without dedicated data engineering capacity to build the fund-data domain layer themselves
- Firms running Databricks who need a purpose-built domain layer that consumes Delta tables and writes governed outputs back
- Organizations that want pre-built fund-data connectors and ontology rather than assembling them on top of a lakehouse
Integration
How PlexiFact + Databricks Work Together
Three deployment patterns we see most often. None of them require you to abandon what you already have.
PlexiFact reads Delta Lake
Use Databricks as the lakehouse: PlexiFact ingests Delta tables, applies fund-data schema and lineage, and exposes governed outputs through connectors. Unity Catalog metadata flows through.
PlexiFact writes Delta Lake
Run PlexiFact as the fund-data layer and land curated, reconciled outputs back into Delta tables for ML, BI, or downstream lakehouse consumers in the same platform.
Databricks SQL as the compute backend
For customers standardized on Databricks, PlexiFact can delegate heavy SQL workloads to Databricks SQL Warehouses or Photon while keeping fund-data orchestration, governance, and connectors in PlexiFact.
Already running Databricks?
Good - we will plug into it. Let us scope the fund-data layer that sits on top of your lakehouse, and we will give you an honest read on which deployment pattern fits.