Sign in

Robert Zych

@zychr.bsky.social
379 followers 180 following 42 posts

Software Engineer @ Raft, Apache Pinot Contributor, lives in Sacramento area.

PostsRepliesMedia
Robert Zych @zychr.bsky.social · 09/12/2025
Your book on functional programming is still on my bookshelf as it’s one of my all time favorites. Looking forward to reading this one.
120
Robert Zych @zychr.bsky.social · 13/07/2025
Which techniques do you use to increase your trust in model generated code? Submitting the same task to multiple models can produce interesting results.
010
Robert Zych @zychr.bsky.social · 08/03/2025
For this demo I generated a 100TB data set (50GB/min) and maintained sub-second query latencies teamraft.com/resources/in...
teamraft.com
Secure & Low-Latency Queries at scale with Raft Data Platform - [R]DP - Raft | Operationalize Data & Agentic AI
Raft is a DefenseTech Product company on a mission to connect humans with data at the edge. One of our products is the Raft Data Platform -[R]DP. We’ve built the [R]DP and operationalized it in severa...
010
Robert Zych @zychr.bsky.social · 07/03/2025
Ale jaja
110
Reposted by Robert Zych
Kaivalya Apte @thegeeknarrator.bsky.social · 02/02/2025
The JVM internals series - #5 is out.. This episode lays out a good foundation for us to go into the internals of each of the amazing GCs and their tradeoffs. Become a member to access the upcoming internals episode.. Meanwhile watch this video (free access until tomorrow). youtu.be/p1Yamdly0QE
041
Robert Zych @zychr.bsky.social · 31/01/2025
Perfect Polish smile. Just need a black and white copy for your passport.
010
Robert Zych @zychr.bsky.social · 25/01/2025
With @apachepinot.bsky.social I can query a 100TB data set and using it’s Star-tree index can get 20ms query latency!
030
Robert Zych @zychr.bsky.social · 17/01/2025
Has anyone used DuckDB for TB or PB scale data sets? Asking for a friend.
120
Robert Zych @zychr.bsky.social · 25/12/2024
I was reading this book recently and stumbled across this handy formula for calculating Kafka partitions: Max(target tp / producer tp, target tp / consumer tp)
020
Robert Zych @zychr.bsky.social · 17/12/2024
@startreedata.bsky.social Thank you and Merry Christmas! My community box arrived last night and with my winnings from the MI competition I recently purchased a pair of skis. #startree
121
Robert Zych @zychr.bsky.social · 11/12/2024
I’m planning to use ST to generate 100TB of data
130
Robert Zych @zychr.bsky.social · 07/12/2024
@apachepinot.bsky.social It's a bit early for xmas, but thanks for all the features! youtu.be/gwcL7O0o6W4?...
youtu.be
Meetup: Apache Pinot Year in Review 2024
YouTube video by StarTree
010
Robert Zych @zychr.bsky.social · 04/12/2024
@apachepinot.bsky.social Welcome!
020
Robert Zych @zychr.bsky.social · 04/12/2024
What is the graph trying to show? Do you have any examples of virtual threads from prod?
000
Robert Zych @zychr.bsky.social · 03/12/2024
Last night, I was exploring thread safety issues with ChatGPT. It only took two attempts to generate a test that exposed the lack of a synchronized block for a section that updates multiple member variables. #java
021
Robert Zych @zychr.bsky.social · 23/11/2024
Assuming refs could handle the new rule it would prevent goals from offside situations, but it would hurt counter attacking.
010
Robert Zych @zychr.bsky.social · 23/11/2024
It’s been a while since we’ve had this discussion. Could you remind me when you would call offsides?
100
Robert Zych @zychr.bsky.social · 22/11/2024
Didn’t we used to say monkey’s tail back in the early days of the internet?
110
Robert Zych @zychr.bsky.social · 19/11/2024
Correct! Thanks for playing! This was based on an example from “Java Concurrency in Practice”.
110
Robert Zych @zychr.bsky.social · 19/11/2024
What does this program output? #java
111
Robert Zych @zychr.bsky.social · 17/11/2024
trod through great compilation Exodus!
010
Robert Zych @zychr.bsky.social · 14/11/2024
Here’s my dog Mocha. She’s 2
110
Robert Zych @zychr.bsky.social · 14/11/2024
@thegeeknarrator.bsky.social thanks for the book recommendation. This one is much easier to read
011
Reposted by Robert Zych
Jaz @jaz.sh · 13/11/2024
90 days ago Bluesky had 6.18M users 30 days ago Bluesky had 10.85M users Tomorrow Bluesky will have 15M users This is absolutely wild, thanks for believing in us y'all
1242559255507
Robert Zych @zychr.bsky.social · 13/11/2024
Coincidentally I just started reading and enjoying “Java Concurrency in Practice”
120
Robert Zych @zychr.bsky.social · 12/11/2024
Let's go! Update your starter packs #dataBS
010
Robert Zych @zychr.bsky.social · 05/11/2024
Great point! Perhaps “performance over precision” is a more appropriate phrase for this trade off. Another example that comes to mind is from Kafka Streams where you define the grace period for windows with late arriving events.
010
Robert Zych @zychr.bsky.social · 05/11/2024
In which cases is performance more important than correctness? Approximations. Like Pinot’s DISTINCTCOUNTHLL which returns an approximate distinct count using HyperLogLog. Does anyone else have a good example?
approximations.like
110
Robert Zych @zychr.bsky.social · 04/11/2024
And your data model can be only as good as the query patterns it was designed to support
110
Robert Zych @zychr.bsky.social · 03/11/2024
Both if possible. But if not, then correctness over performance. And all other key benefits of the Responsive platform itself (above and beyond Kafka Streams)
010
Robert Zych @zychr.bsky.social · 03/11/2024
How does Responsive support calls to external services while processing the stream? What are the current and future benefits of using SlateDB?
120
Robert Zych @zychr.bsky.social · 03/11/2024
Thanks for the refresher! Now I wonder how the false positive probability (fpp) parameter works in Pinot. docs.pinot.apache.org/basics/index...
docs.pinot.apache.org
Bloom filter | Apache Pinot Docs
This page describes configuring the Bloom filter for Apache Pinot
100
Robert Zych @zychr.bsky.social · 03/11/2024
I meant “hand” not “handheld”.
000
Robert Zych @zychr.bsky.social · 03/11/2024
IIRC, Bloom filters are a probabilistic data structure. Meaning that they return a maybe yes or definitely no result. In this case it’s a maybe yes because of the “handheld” part of the query, correct?
200
Robert Zych @zychr.bsky.social · 02/11/2024
Better, but I prefer Kaivalya’s. What’s the final result and why?
000
Robert Zych @zychr.bsky.social · 02/11/2024
I’m a supporter once again! Thank you for your service 🫡
110
Robert Zych @zychr.bsky.social · 02/11/2024
Having the lines going through the bit indices helps to explain the mappings. Though it’s still not clear what the result of the lookup is.
110
Robert Zych @zychr.bsky.social · 02/11/2024
I’d like to understand what you’re showing here but would need to see a more straightforward example first. Are you still accepting financial contributions?
100
Robert Zych @zychr.bsky.social · 01/11/2024
When it comes new data systems Rust appears to be winning for now.
010
Robert Zych @zychr.bsky.social · 01/11/2024
Yes, and contributing to the libraries can take your education to a new level
010
Robert Zych @zychr.bsky.social · 31/10/2024
To make such a big change after so many years is inspiring. I switched from C# (10yrs) to Python (6yrs) to Java (1yr and counting)
040
Robert Zych @zychr.bsky.social · 29/10/2024
Working on JSON predicate pushdown for the Trino-Pinot connector: contains((1,2,3), json_extract_scalar(json, '$.key`)) -> json_match(json, '$.key in (1,2,3)') json_array_contains(json_extract(json, '$.array'), value) -> json_match(json, '$.array[*] = value') json_match uses Pinot's JSON index
030
Robert Zych @zychr.bsky.social · 27/10/2024
Interest, discipline, and potential.
010
Robert Zych @zychr.bsky.social · 08/08/2023
It’s a good sign when a friend surprisingly welcomes you on your 1st day at a new job. Thanks @nbuesing.bsky.social !
240