People, Technology, Culture: How to make Data Science work
summary
Large enterprise data strategy argues for cutting platform complexity, moving legacy workloads from on‑prem to IaaS/PaaS and aiming for data science as a service so teams focus on models not ops. Claims a 20–80x cost gap between enterprise data warehouses and open source stacks, uses Hadoop/Spark/Kafka/H2O/Jupyter/TensorFlow, adopts a Lambda architecture, targets 4–6 week iteration cycles, builds an Explorer data‑visual tool on GitHub and a GDPR‑compliant mobility insights product used for city event planning, and enforces privacy and governance by design.
AI-written summary · about
chapters (19)
- 0:00Start
- 1:47Agenda
- 2:35A1 Telekom Austria Group at a glance
- 2:56About me
- 4:12Thales of Milet
- 6:25What do you want to do? (RAISE YOUR HANDS)
- 11:45Big Data & Analytics is not about proprietary Databases & Services
- 13:23Jon Snow fighting Cholera with Statistics
- 14:52So - should we be afraid of AI?
- 15:28Jump off the bridge!!
- 16:09Who should do Big Data & Advanced Analytics in the enterprise?
- 19:13Explore Package
- 22:50Our solution is fully in line with the data protection law - Data ananymization
- 28:33What is necessary to run Data Science & Machine Learning at scale?
- 29:25The Datalake risk...
- 30:20How will/did we solve this? Challenges?
- 31:49This is all about the Datalake - Layers. Based on Lambda, best of breed
- 33:56Agile Data Science
- 37:33Data Governance framework for Group Data Lake
related talks
41mOn Data: Sky above, sand below, peace within
08mThe Democratization of Data Science
51mFrom Data Confusion to Data Intelligence
13mEnterprise-Level Data Science Success
42mWhy Big Data Is Not Data Science
12mThe Analytical Imperative
58mGrowth & change of careers, organizations & responsibility in data science
23mData science as a team sport