FundingNov 2023
Medovac raises $85M Series B led by Andreessen Horowitz
The round accelerates research into causal inference and the Atlas semantic layer.
Autonomous data science, in production
Ship decisions, not dashboards.
Medovac turns raw operational data into governed, production-grade intelligence. Our platform automates the data science lifecycle: ingestion, feature engineering, modeling, causal analysis, and monitoring, so analytics teams move from question to decision in hours.
SOC 2 Type IIISO 27001HIPAA readyGDPR
Live platform telemetry
streamingrecords/s
48,213
p50 latency
1.9s
models live
1,284
Built with and trusted alongside
AWSNVIDIAGoogle CloudMicrosoft AzureSnowflakeDatabricksdbt LabsConfluentAWSNVIDIAGoogle CloudMicrosoft AzureSnowflakeDatabricksdbt LabsConfluent
The platform
Five engines, one governed data plane
Each engine automates a stage of the data science lifecycle. Together they take you from raw records to a decision you can defend.
Medovac Signal
Automated feature intelligence
Signal profiles your warehouse, proposes candidate features, and promotes the ones that survive statistical validation into a governed feature store.
Learn moreMedovac Forecast
Probabilistic forecasting
Forecast blends temporal fusion transformers with classical statistical baselines and returns full predictive distributions, not point estimates.
Learn moreMedovac Sentinel
Anomaly and drift monitoring
Sentinel monitors data quality and model health with unsupervised anomaly detection and adaptive thresholds tuned to each metric's own history.
Learn moreMedovac Causal
Causal inference and experimentation
Causal estimates treatment effects from experiments and observational data using doubly robust estimators and synthetic controls.
Learn moreMedovac Atlas
Semantic analytics layer
Atlas is a semantic layer and retrieval-augmented analyst that translates natural language into governed queries against certified metrics.
Learn moreOne platform, not five tools
Shared lineage, shared governance, shared security. Adopt one engine or all five: they speak the same data plane.
See how it fits togetherBy the numbers
Rigor that holds up under load
Aggregated across Medovac production deployments in regulated industries.
0.0T
Records processed monthly
across customer data planes
0%
Median analyst time saved
on recurring modeling work
0+
Enterprise deployments
in regulated industries
0.00%
Platform uptime
trailing 12 months
Measurable impact
What autonomous data science looks like in the numbers
Real patterns from Medovac deployments: faster delivery, lower error, and throughput that scales with your business.
Time from raw data to governed production
Traditional data science workflow47 days
notebook to governed production
In-house platform build21 days
Medovac autonomous platform3 days
connect, automate, ship
Median across comparable enterprise projects. Lower is better.
Monthly throughput growth
Records processed per month, in billions, trailing 12 months.
Forecast error by approach
Seasonal-naive baseline18% MAPE
Classical (ARIMA / ETS)14% MAPE
Medovac Forecast9% MAPE
temporal fusion + conformal intervals
Mean absolute percentage error on held-out backtests. Lower is better.
Outcomes customers report
0%
Analyst time reclaimed
vs. manual pipelines
0%
Lower forecast error
vs. prior baseline
0%
Fewer false alerts
with adaptive thresholds
0%
Fewer ad-hoc requests
answered by Atlas
How it works
From raw records to a defensible decision
Medovac closes the last mile of analytics. Every stage is automated, governed, and observable, so your team spends its judgment where it matters.
- 01
Connect your lakehouse
Read-only connectors profile every table and build a governed semantic model.
- 02
Automate the lifecycle
Signal engineers features, Forecast and Causal model outcomes, Sentinel monitors health.
- 03
Ask and act
Atlas answers plain-language questions with lineage, and recommendations flow to where work happens.
Live platform telemetry
streamingrecords/s
48,213
p50 latency
1.9s
models live
1,284
Representative telemetry from a live Medovac data plane.
Newsroom
What's new at Medovac
ProductSep 2024
Atlas brings governed natural-language analytics to every team
Ask questions in plain language and get answers grounded in certified metrics with full lineage.
PartnershipJun 2024
Medovac joins the NVIDIA Inception program
Deepening GPU-accelerated training and inference across the forecasting and causal engines.
See Medovac on your own data
Book a technical walkthrough with our field data science team. We will connect a sample of your data and show governed, production-grade intelligence in under an hour.