InsurTech
Data platform for a GenAI product
An InsurTech was building a GenAI product. We designed the data platform and vector store on AWS, with retrieval fast enough for production traffic.
Data Engineering
Pipelines, data platforms and MLOps on AWS, Azure, GCP and Databricks.
We build data pipelines, migrate data to the cloud, set up lakehouses and the MLOps tooling to deploy models. We work on AWS, Azure, Google Cloud, Databricks and Swiss providers such as Infomaniak.
The problem
Data often sits in old systems, in several clouds or in spreadsheets, with quality problems nobody has time to fix.
Models and dashboards built on top of it break regularly.
How we work
- Data Landscape Assessment
Map where your data lives, how good it is, and what to fix first.
- Architecture & Platform Design
Choose the platform and design an architecture that will still fit in a few years.
- Legacy Modernization
Connect or replace legacy systems so their data can be used.
- Data Transformation
Build the pipelines that clean, join and load the data, in the cloud or on-premises.
- Enablement & Handover
Train your team and hand over code, documentation and runbooks.
What you get
- Cloud data migration
Your data on AWS, Azure, GCP or Databricks, structured and documented.
- Production-grade pipelines
Tested pipelines with monitoring and alerting.
- MLOps infrastructure
Tooling to deploy, monitor and retrain models.
- Full-stack ML tooling
Everything from data ingestion to the interface people use.
Projects
InsurTech
Data lakehouse on Databricks
An InsurTech had its data spread across several systems. We built a lakehouse on Databricks: architecture, pipelines, transformations and MLOps. Their data scientists now work from one source.
Industrial Robotics
Cross-cloud analytics pipeline
An industrial robotics company kept machine data in three cloud environments. We built a pipeline across AWS, Azure and Databricks that feeds one analytics layer for production monitoring.
FinTech
LLM deployment on Swiss Cloud
A regulated FinTech needed a language model that keeps all data in Switzerland. We deployed the model, the data pipeline and the access controls on Infomaniak Swiss Cloud.
Retail
Competitor pricing intelligence
A retailer set prices using competitor data that was a week old. We built a pipeline on Google Cloud that collects and processes competitor prices every day.
Write to Fredrik directly.
Fredrik Moeschlin CEO & Founder