Eran Levy's Avatar

Eran Levy

@levyeran

Technology, Business and in between

26
Followers
220
Following
9
Posts
24.11.2023
Joined
Posts Following

Latest posts by Eran Levy @levyeran

Preview
Microsoft CPO says she 'fundamentally' disagrees with the idea that people shouldn't study computer science The Microsoft CPO says it's possible software engineers will be viewed as "software operators" in the future, but the role is unlikely to disappear.

www.businessinsider.com/microsoft-cp...

Just thought about this recently, what will happen with the CS studies and I like the way she is explaining this shift - higher level programming and operations, you will still need to understand the underlying..

24.05.2025 14:53 ๐Ÿ‘ 1 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
The Architect Role Evolution: A Perspective from My Own Journey The traditional architect role no longer fits modern software development and companies started adjusting

Iโ€™ve spent years in architecture roles and I observe the natural change in the last years - the traditional "architect" role no longer fits modern software development and companies started adjusting.
levyeran.medium.com/the-architec...

09.03.2025 17:41 ๐Ÿ‘ 1 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
The End of Programming as We Know It

Very nice post by Tim O'Reilly - We are at a historic moment and Im really curious to see how AI will shape the future of software engineering. Im one of the people that dont think that our job as engineers will be obsolete- rather they will evolve

www.oreilly.com/radar/the-en...

24.02.2025 04:20 ๐Ÿ‘ 1 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
How Uber Uses Rayยฎ to Optimize the Rides Business Computational efficiency is a significant challenge when scaling solutions to a marketplace as large and as complex as Uber. The running and tuning of the Uber rides business relies on substantial numbers of machine learning models and optimization algorithms. Luckily, many pieces of the system can be treated in parallel. Rayยฎ is a general compute engine for Pythonยฎ thatโ€™s designed for ML, AI, and other algorithmic workloads. We describe how Uber has adopted Ray to enable mission-critical systems. Using Uberโ€™s mobility marketplace allocation tuning system as an example, we found performance improvements of up to 40 times that unlocked new capabilities. It also improved developer productivity by increasing iteration speed, reducing incident mitigation time, and lowering code complexity.

Very nice use case for utilizing Ray at Uber
www.uber.com/en-JO/blog/h...

11.01.2025 18:27 ๐Ÿ‘ 0 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0

Hilarious;)

12.12.2024 05:09 ๐Ÿ‘ 1 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0

The new @awscloud.bsky.social S3 #iceberg tables (while still seems broken) theoretically solve 3 challenges: 1.metadata-solve today complexities with data retention, tag, security.. 2.Performance-a good promise, 3. Unify all under the same umbrella? competing with #azure fabric and others?

05.12.2024 12:45 ๐Ÿ‘ 0 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
New Amazon S3 Tables: Storage optimized for analytics workloads | Amazon Web Services Amazon S3 Tables optimize tabular data storage (like transactions and sensor readings) in Apache Iceberg, enabling high-performance, low-cost queries using Athena, EMR, and Spark.

aws.amazon.com/blogs/aws/ne... now AWS fully supports Iceberg with a new type of S3 table buckets with all the data maintenance built in, great news!

03.12.2024 21:26 ๐Ÿ‘ 1 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
Agodaโ€™s Unconventional Client-First Transition from a GraphQL Monolith to Microservices Agoda recently described their innovative approach to transitioning from a monolithic GraphQL API to a microservices architecture. Unlike traditional methods focusing on breaking down server-side comp...

www.infoq.com/news/2024/11... very nice approach and out of the box thinking

27.11.2024 06:10 ๐Ÿ‘ 0 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
An Experiment in Interactive Education: How LLMs Are Transforming the Way Our Kids Learn Iโ€™m fascinated by how LLMs are starting to impact education, especially for our children. The transformation is already underwayโ€ฆ

Some experiment done with my kids and @OpenAI #ChatGPT ;) and some discovery after reading my daughter's @MSFTCopilot chat history. Fascinated by how LLMs are starting to impact education, especially for our children
levyeran.medium.com/an-experimen...

10.11.2024 09:41 ๐Ÿ‘ 1 ๐Ÿ” 0 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Illustration of how a Bloom filter works. At the top are three keys that exist in the Bloom filter; for each we show the hash as an 18-bit binary number broken up into 3 groups of 6 bits. In the middle are 64 boxes each representing one bit. The boxes are numbered to 0 to 63, and each of the 6-bit values is interpreted as an index into the array of boxes. There are arrows from each 6-bit value to the corresponding box, and the boxes with an incoming arrow have a 1, whereas the others have a 0. At the bottom is a key being queried, which also has an 18-bit hash, and the arrows from the three 6-bit parts of that hash also point at the appropriate boxes.

Illustration of how a Bloom filter works. At the top are three keys that exist in the Bloom filter; for each we show the hash as an 18-bit binary number broken up into 3 groups of 6 bits. In the middle are 64 boxes each representing one bit. The boxes are numbered to 0 to 63, and each of the 6-bit values is interpreted as an index into the array of boxes. There are arrows from each 6-bit value to the corresponding box, and the boxes with an incoming arrow have a 1, whereas the others have a 0. At the bottom is a key being queried, which also has an 18-bit hash, and the arrows from the three 6-bit parts of that hash also point at the appropriate boxes.

For the second edition of Designing Data-Intensive Applications we want to add a brief explanation of what a Bloom filter is (even though it's quite a well-known concept, we don't want to assume readers already know it). Here's my attempt at showing it diagramatically. What do you think?

02.11.2024 14:36 ๐Ÿ‘ 159 ๐Ÿ” 21 ๐Ÿ’ฌ 21 ๐Ÿ“Œ 4