I’ve had a brilliant 2-and-a-bit years with Cloudera, but in a surprising turn of events, I’ve decided it’s time for something a little bit different. I left Cloudera at the at the end of May, and in June I will be starting at Tinybird in their fledgling UK team.
Tinybird is a new SaaS company, with around 40 people, originally formed in Spain by a bunch of pretty brainy people. They’re building a platform that is the embodiment of which I have been advocating for several years now – the productisation of real-time data.
Streaming & real-time are similar, but different beats, and both are pretty hard – but new players, like Tinybird, are starting to make it easier. We’re at the stage where many companies have adopted Kafka/Kineses/Pulsar or some kind of messages platform; we’ve gotten good at building streams, but it turns out that its still pretty hard to actually do something with it.
Streaming analytics (or stream processing) is one such something, and its a very exciting space that is growing in populatiry, but still has a high bar to entry – Apache Flink is the engine de jour for this, and while it has had a lot of success, many find it incredibly difficult to be productive with it (shoutout to Decodable who are making Flink more accessible). There is also ksqlDB from Confluent, but as they say, when you only have a hammer, everything looks like a nail. Anyway, streaming analytics is the idea that we query data in-flight, allowing us to find patterns over windows of time and gain insight as the data is generated. It’s very cool, it’s very hard, and it has tremendous value.
Another something is what is often just called ‘real-time data’. In real-time data, we are less concerned with running constant analytics over a stream, and instead we want to make our data available to our applications as fast as possible. Typically, these are interactive applications with strict latency, freshness and concurrency requirements. Think of the ‘who viewed my profile’ notifications on LinkedIn – this is real time data made available to millions of active users.
Tinybird is building a platform for the latter – how do we make our data, wherever it comes from, available wherever it needs to be, as fast as possible – and do that with minimal cognitive load on developers. I think Tinybird has cracked it with a great technical product and a fantastic developer experience.
I’m excited to get started with Tinybird very soon, where I’ll be continuing in a technical Customer Success role, getting my hands dirty and making the next generation of real-time data products.