Speaker: George Nikolaropoulos, SVP of Engineering, GWI
What does your company do and what is your role there?
George Nikolaropoulos: At GWI, we build technology that gives people clarity over their audiences in the moment they need it. We're the leading audience targeting company for the global marketing industry and we show our customers how their consumers think. We've been delivering game changing insights to all the agencies, media companies and brands, you all know and love since 2009. I'm Geroge Nikolaropoulos, Senior Vice President of Engineering at the GWI and I'm overseeing the teams of DevOps, QA and engineering.
Why did you choose Google Cloud?
George Nikolaropoulos: Google Cloud was for us, the obvious choices are highly available and highly containerized solution meaning nothing but the best kubernetes out there. So for us, it was always GKE but it wasn't only that we needed a platform that would allow our teams to collaborate and deliver effectively to our customers. Hence now we have one platform where our BI and data science and engineering team can work together and become extremely productive. DevOps team also loves the simplicity of the UI and the capabilities of the tools they find in there that allow them to become more productive daily. Last but not least I need to mention that we've built some extremely good relationships with the people at Google and our third party vendors. Something that's beyond anything I've ever experienced with any other vendor. And for us building long lasting relationship with mutual trust is really important
What have you achieved using Google Cloud solutions?
George Nikolaropoulos: We use a wide range of offerings within Google Cloud. From our BI teams using BigQuery and Data Studio to our data science teams using the AI services and lately Vertex AI to the engineering team using well, almost everything, the biggest challenge we had as a scale up was to continue to scale with a cost effective fashion. Moving to Google Cloud allowed us to reduce by 20% the cost of our infrastructure and also enabled us to continue to grow with a sublinear cost. Heavily using auto-scaling capabilities within compute engine allow us to hit an uptime of 99% more consistently. Last but not least we managed to reduce our mean time to recover by 50%, as with all the logging and monitoring tools within Google Cloud, discovering the root cause of a production issue is now a walk in the park.
What are your future plans for working with Google Cloud?
George Nikolaropoulos: As we continue to grow and scale at GWI I think the next obvious step is going serverless. So things like Cloud Run and Cloud Functions are always in our radar. Data privacy is also very close to our heart. So things like data clean rooms and other technologies that promote data anonymity and privacy is something that we're always looking for. But remember, we are an audience analytics company. So what we actually do is providing astonishing insights for the consumers of our customers as fast as we can and what a better use for AI and LLMs. So really in the next coming months, I cannot wait to see how my teams are going to push the boundaries of the PaLM2 models like unicorn and bison.
