A large U.S. telecom provider’s Construction & Engineering (C&E) organization launched a multi-year initiative to modernize its analytics ecosystem and shift from on-prem, warehouse-centric reporting to a cloud-based, Data Mesh-aligned model. Cadmus played a critical role in migrating legacy data (10+ Petabytes) and processes; re-engineering complex datastores to Snowflake; ETL from SQL Server/SSIS to scalable Palantir Foundry pipelines; and transforming the user experience from static “canned” reports to near-real-time, self-service analytics.
The program combined deep engineering (distributed transformations, orchestration, performance tuning, and operational hardening) with change management (training, adoption enablement, and a pragmatic reporting transition). Most recently, we extended the platform with AI-powered capabilities, including a GenAI-enabled conversational analytics chatbot using Palantir AIP, now in production and expanding into operational workflows.
The C&E organization’s legacy environment had grown over time and included:
This model constrained agility, increased operational burden, and limited the organization’s ability to ask new questions quickly. The target state was a modern platform that could support scalable processing, improved governance, decentralized ownership, and self-service insights, moving teams from “request a report” to “answer it yourself,” with fresher data.
Cadmus aligned delivery to the C&E organization’s modernization goals:
This model constrained agility, increased operational burden, and limited the organization’s ability to ask new questions quickly.
A defining constraint early in the program was the lack of meaningful vendor support during implementation. When we started in 2020, there was limited structured training and limited hands-on guidance for how to translate a complex legacy data environment into Palantir Foundry at scale.
A defining constraint early in the program was the lack of meaningful vendor support during implementation.
We closed this gap through deliberate internal capability building:
This approach reduced vendor dependency and created a sustainable internal delivery capability for the client.
This approach reduced vendor dependency and created a sustainable internal delivery capability for the client.
We migrated application data from SQL Server to Snowflake, so our ingestion approach used a Snowflake-replicated copy of the application database as the primary source. Snowflake served as an intermediary/source system for multiple consumers, while Foundry became the core platform for processing and self-service analytics enablement.
This was not a lift-and-shift. It was a re-platforming effort that changed how transformations executed, scaled, and were operated.
Key actions included:
The largest technical lift was re-engineering legacy transformation logic originally implemented in SSIS and SQL into a cloud-scale execution model. We rebuilt transformations primarily using PySpark, and leveraged Snowpark/SnowSQL where Snowflake-native processing was the best fit.
Core engineering outcomes:
This was not a lift-and-shift. It was a re-platforming effort that changed how transformations executed, scaled, and were operated.
With the investment in modern data architecture, the client upgraded data operations standards. A Data Mesh approach provided the framework to deliver features quickly at scale by treating curated datasets as governed, reusable data products owned by domain teams.
As part of the Data Mesh modernization, Cadmus built foundational capabilities that made the platform scalable, trustworthy, and easier to adopt:
These capabilities reduced operational drag and made self-service analytics viable at scale.
A key objective was moving from once-per-day reporting to near real-time analytics. We designed ingestion and transformation pipelines to support faster refresh cycles and operational use cases, including:
In a high-volume telecom environment, even small delays or instability quickly erode confidence.
To accelerate delivery and iteration speed, we leaned on Foundry’s developer tooling:
To improve confidence and operational maturity, we leveraged:
Once end-to-end flows were established, we focused heavily on performance and stability:
In a high-volume telecom environment, even small delays or instability quickly erode confidence. We focused on performance and reliability, so the platform stayed fast, dependable, and trusted for day-to-day decision making.
The client’s intent was not only data processing modernization, but also self-service analytics adoption. Foundry’s Contour Dashboards were central to shifting behavior.
Our adoption approach:
Contour adoption faced early resistance, especially from teams used to familiar reporting tools. We avoided disruption by introducing Power BI as a transitional layer, providing continuity for key stakeholders while training and adoption matured. As self-service usage increased, reliance on Power BI declined and the organization shifted to Foundry-native analytics.
To improve usability and analytic flexibility, we also used Quiver alongside Contour. This strengthened self-service by giving users more intuitive exploration and interaction patterns while maintaining governance.
Result: users gained control without requiring developers for every new insight, and teams moved closer to real-time, decision-support analytics.
Users gained control without requiring developers for every new insight, and teams moved closer to real-time, decision-support analytics.
In addition to analytics, the client used Foundry for small-scale operational workflows that benefit from governance and traceability but are not high-concurrency transactional systems, including:
These solutions fed downstream systems while maintaining auditability and controlled updates.
Most recently, we implemented conversational analytics using Palantir AIP:
We are actively expanding AIP to integrate conversational analytics and predictive insights into operational workflows, especially around project health monitoring, so insights show up where work happens rather than staying trapped in dashboards.
This engagement delivered transformation across platform, operations, and user behavior:
Innovation Outcomes
This program reflects the full arc of enterprise modernization done the hard way—and done right:
