Oliver Libuda’s Post

An inspiring journey with DSFP 4.0 fellows as they learn, grow, and transform into the next generation of data scientists. Watch highlights of their in-class activities, teamwork, and passion for learning and making a difference through data. Watch Full Video: https://lnkd.in/ddak8EEi #ClickOnKadunaDSFP #DataScience

🚀 Excited to share that in just a few days I'll be speaking at the Open Data Science Conference (ODSC) West! I’ll be presenting a session titled “Explainability Explained: From Beta Coefficients to SHAPly Values” – diving into the journey from traditional regression insights to modern machine learning interpretability techniques. Looking forward to connecting with fellow data science enthusiasts and sharing new perspectives on explainability. See you there! #ODSC #DataScience #Explainability #MachineLearning #ModelTransparency

Introducing FeatEng, a benchmark designed to evaluate LLMs on one of the most knowledge-intensive tasks in data science: generating feature engineering code. Learn about the intersection of LLMs and data science, including: - A new approach to benchmarking LLMs that integrates skills across domains - Key strategies LLMs use for advanced feature engineering - Insights into which models excel and where they fall short Full details: https://okt.to/DBjx05

We're only a few days away from our live information session! Join us April 18 at 12 p.m. PDT to get answers to your questions regarding our online M.S. in Engineering areas of study in data science engineering and engineering management. It's not too late to secure your spot for the webinar. Use this link to register: https://bit.ly/3TIrNpy #EngineeringSchool #DataScience #EngineeringManagement

🎄 Day 4 of the DataExpert.io Bootcamp! I took a break yesterday to enjoy a wonderful evening with my wife at our town’s Christmas tree lighting 🎅🌟—but I’m back on track today! I wrapped up Day 2 lecture and lab solo with no issues and learned some key data engineering concepts: 🔑 Idempotency: What it is, what causes it, and how it impacts pipelines. 🔑 Slowly Changing Dimensions (SCDs): Explored the different types of SCDs and two approaches to building pipelines to handle them: Full data loads: Processes all data but can be expensive due to window functions on large datasets. Incremental loads: A more efficient approach that breaks data into CTEs (historical, unchanged, changed, and new records) and unions them into a single table. It’s more complex to build but significantly reduces compute cost—processing up to 20x less data! I also learned about potential pitfalls of incremental pipelines, like filtering out data when columns are null, and how to ensure assumptions about data integrity hold true. Feeling confident after today’s progress and excited to dive into more labs! 📢 To everyone tackling similar challenges, what’s been your biggest takeaway so far in building pipelines? Catch me Monday morning on Twitch for more streams: https://lnkd.in/ebWqpRXH #DataEngineering #BootcampJourney #IncrementalLoads #SCD #TwitchStreaming #LearningTogether

Just wrapped up Day 2 of the Data Engineering Bootcamp! This session was all about Slowly Changing Dimensions (SCD), with a deep dive into the pitfalls of SCD-1 and why it’s best to avoid it. Zach is picking up a fight in Data Engineering world by openly declaring his dislike for SCD-1 (for all the right reasons). Well, taking a stand always stands out. ✨ Ended the lesson with some thought-provoking wisdom: “You don’t have to be caught up with the idea that every pipeline you build is a Ferrari. What’s happening is you are wasting time on a marginal value which could have been used for something more productive. Always take into account the opportunity cost” Get the short notes of the session here 🔽

Week 7 of Summer School is almost here! 🚀 We're excited for tomorrow's class, which will feature an engaging session on data methodology. In case you missed it, check out the clip 🎥 below from last week’s session with our new strategic data skills partner, Corndel, where Thomas Neal, Avtar Singh Boparai FCCA/CMgr and Stuart Bevis led an interactive session on the crucial need to train employees in technical skills 🤝 #DataMethodology #DataTransformation #DataCommunity

I’m completely blown away by the knowledge bombs 💣 Zach Wilson is dropping — and this is just Day 1 of his DataExpert.io bootcamp! I knew different data consumers required different approaches, but the concept of master data and how it fits into OLTP and OLAP workflows was something completely new to me. The highlight was the cumulative table design — for some reason, it weirdly reminded me of the GIF protocol (still figuring out why 🤔). Seeing how historical data can be seamlessly preserved with a FULL OUTER JOIN and smart structuring — and then trying it hands-on in the first lab — made me realize just how powerful this design truly is! Can’t wait to see what the rest of the BootCamp holds. #DataEngineering #DataModelling #CareerDevelopment #LearningInPublic #BigData #TechSkills #ProfessionalGrowth

This is going to be a very educational and potentially eye-opening webinar, especially for those of us who always loved the idea of semantic layers and knowledge graphs, but could never quite find the compelling business case for them. GenAI and LLM change that equation significantly. I highly recommend tuning into this webinar to find out why and how! Fluree Eliud Polanco

I highly recommend joining this webinar. If you are not thinking semantic layer, then you will not be able to keep up with the proliferation of data that will be coming at you. It will seem unmanageable. Only through automation will you be able to be successful and the semantic data layer will be your key to success. Definitely worth your time.