Ulrike Hoffmann-Burchardi, Ali Ghodsi


Sync Scroll

The easy way to rediscover your favorite moments of the talk.
Time Topic
0:05 Welcome by Steffi Czerny
3:30 Ulrike Hoffmann-Burchardi introduces Databricks and Ali Ghodsi
4:25 Ali Ghodsi on his childhood in Sweden and early fascination with coding and computers.
6:10 The origins of Databricks: How Ali Ghodsi and Ion Stoica met at U.C. Berkeley.
7:30 Developing Apache Spark. “This is almost like 1960s: There’s a new computer, and the new computer is the data center. It’s the cloud.” (Ali Ghodsi)
9:05 Artificial intelligence in data centers: How a Netflix competition sparked the making of Spark.
11:35 “We want the whole world to take this and run with it.” Turning research into a product.
12:15 With a little help from Ben Horowitz and a16z…
13:35 The challenge: making people use Spark instead of Hadoop.
15:05 An event in 2015 changes everything: the “sort contest”.
16:40 Suddenly everybody is talking about Spark – but “nobody knew that Databricks is the company behind Spark.”
18:00 The next challenge: earning money. “We succeeded in having impact in the world with open source project. We’re not business guys.”
20:10 Ali Ghodsi becomes CEO, benefits from his “data driven” background in academics.
21:45 “Most researchers are pretty good marketers. You have to explain the research results and convince other researchers who are very skeptical…”
24:00 Three pivots in 2016 that make all the difference.
24:45 The Delta project: addressing the needs of enterprises.
26:30 The hardest part about AI? Making the data useful for smart algorithms.
27:30 Why Databricks takes an integrated approach: “I think our secret sauce is that we combine the data processing and AI in one platform.”
28:15 Data lakes, data swamps, and the Lakehouse concept.
31:30 Why purified data is better data.
33:00 Regeneron, DNA databases, drug discovery: How companies use Databricks to battle Covid-19.
35:00 Data analytics and AI in predictive maintenance. “We tell them in advance, what’s the probability that some device will break down.”
37:30 Better usability with the help of automatic machine learning. “We automate, under the hood, a lot of the things that the data scientists today do manually.”
39:10 Managing the whole lifecycle of data, from preparation to production and back.
42:00 The advantages of open source.
43:00 Advice to CIOs: “Build with AI in mind. Don’t think about AI and data science as an afterthought.”
44:40 The right cloud computing strategy.
48:05 Advantages of data streaming.
51:15 Understanding unstructured data, nuances, emotion? Score one for human intelligence.
54:30 Data analytics and AI trends in 2021.
56:50 Advice to the young: “Double down on something and become worldclass in it.”
1:01:40 Wrap-up.

More DLD Sync

DLD Sync, biological manufacturing
If the 20th century was the Age of Physics, the central technology of the 21st century will be biology. Despite marvelous advances in areas like health care, we have only seen a tiny part of what this powerful technology can do. Multiple…
Herman Narula, Ben Vickers, DLD Sync
Our future in virtual worlds is closer than you think. They will impact our economy, how we work and how we live. The coming revolution will begin to recalibrate how we think of ourselves and our identities, offering a vast opportunity to…
linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram