The Big Interview: GEICO VP, Infrastructure, Rebecca Weekly

Insurance company, GEICO is sticking to its open-source guns with a sweeping infrastructure buildout, as it continues to repatriate many cloud workloads and massively modernise its entire network architecture.

A core focus is avoiding lock-in. An early lift-and-shift to public cloud saw the company saddled with a $300 million annual cloud bill. 

The company has now populated two new data centres with Open Compute Project (OCP) hardware and open-source software, having sourced modular servers that are extensible base platforms for compute, storage, or networking across its growing number of ORV3 racks.

That shift has seen GEICO realise a claimed 50% plus reduction in unit cost for compute per core and a 60% plus reduction in unit cost of storage per gigabyte “compared to what we pay our public cloud providers today and the OEM systems we used prior” Distinguished Engineer Sahid Jaffa said in an OCP whitepaper published in 2025. 

But there’s a lot of ongoing work to do. That means that in an otherwise challenging market for engineers, GEICO also has jobs aplenty available, from AI programme managers, to data centre hardware specialists

Infrastructure VP Rebecca Weekly, for example, is hiring widely across data centre management, network modernisation, service mesh and beyond. Many represent an opportunity to build skills in a company that is taking a novel “roll its own” approach at enterprise scale. 

We sat down to chat with her. 

Breaking a legacy network

The infrastructure modernisation programme doesn’t just span repatriating applications and building its own racks for compute and storage. Networks make for a compelling case in point.

Weekly told The Stack that she inherited a severely underprovisioned network architecture after GEICO got “sold on a proprietary, overpriced, custom MPLS cloud that created additional latency and lack of flexibility in connecting new offices and sites,” as she puts it bluntly.