Sign in

Antón

@anton.galiglobal.com
687 followers 109 following 251 posts

Senior Software Engineer who's enthusiastic with data systems, agentic engineering and programming 🦀. Plus, I'm a big fan of reducing my carbon footprint 🌱

PostsRepliesMedia
Antón @anton.galiglobal.com · 26/03/2026
I'm using openobserve locally to monitor code agents and some internal tools. It works pretty well, it's a single binary and it has plenty of features (but it's not bloated yet)
040
Antón @anton.galiglobal.com · 19/03/2026
This is very cool!
050
Antón @anton.galiglobal.com · 12/03/2026
Pro tips for debugging compilation times in GitHub workflows: - Use `cargo build --timings` to generate an HTML report. - Use the new GitHub feature to attach it as an artifact and browse it directly in GitHub. - Update Swatinem/rust-cache and check the logs github.blog/changelog/20...
040
Antón @anton.galiglobal.com · 07/03/2026
I attended with my daughter the Oracle4Girls event in the local computer science university today. We need more events like this. Huge thanks to all the volunteers who made it possible ❤️
020
Antón @anton.galiglobal.com · 06/03/2026
Next, I did it with one in C, highly optimized. It required more work, the performance is similar... but I solved a concurrency bug in the C implementation antonmry.github.io/FastFiRSt/be...
110
Antón @anton.galiglobal.com · 06/03/2026
Took a Python DNA sequencing utility and rewrote it in Rust. Outcome: ~50× faster than CPython and ~20× faster than PyPy… in one shot antonmry.github.io/FastFiRSt/be...
150
Antón @anton.galiglobal.com · 05/03/2026
I have signed up some days ago but I didn't receive the newsletter and I don't have access to the link. Is that correct? No problem at all if it is, I just wanted to double check in case something isn't working as expected
110
Antón @anton.galiglobal.com · 15/01/2026
Testing opencode with LimaVM as a sandbox. Being able to run `opencode serve` in the VM and `opencode attach` from the host is very handy. Good news: it still works with my Claude Code subscription
010
Antón @anton.galiglobal.com · 28/12/2025
Mandatory infographic. I like how well Nano Banana Pro is able to identify the main ideas in my posts
000
Antón @anton.galiglobal.com · 27/11/2025
My zone 2 run was quite fun today
070
Antón @anton.galiglobal.com · 24/11/2025
I've discovered stacked panes in zellij (a rust version of tmux) this weekend and I can't live without them anymore
1121
Antón @anton.galiglobal.com · 23/11/2025
Nano Banana Pro is amazing 🤩
100
Antón @anton.galiglobal.com · 14/11/2025
Closing an important phase today after three months working on projects at @fic-udc.bsky.social at @udc.gal. Grateful to the DB Lab and to the Green IA & algs chair for an inspiring journey. I’ll keep collaborating with the university but there are new exciting things ahead!
020
Antón @anton.galiglobal.com · 12/11/2025
I had a lot of fun with this session ❤️ TIL IO. println() in Java
010
Antón @anton.galiglobal.com · 11/11/2025
My DS213J is 11 years old (and still working!). Time for an upgrade
030
Antón @anton.galiglobal.com · 10/11/2025
My notes for the session, I don't use slides for this one but a combination of Markdown in Neovim and jshell: github.com/antonmry/jav...
100
Antón @anton.galiglobal.com · 01/11/2025
I thought my next post would be something different, but this project ended up being enjoyable enough that it deserved a write up. “Building LocalFirstPacer: a small running tool for someone you care about” www.galiglobal.com/blog/2025/20...
020
Antón @anton.galiglobal.com · 01/11/2025
I may not be using Java often right now, but talking about the latest features is always worthwhile, and doing so in the local Java User Group is a bonus www.linkedin.com/feed/update/...
130
Antón @anton.galiglobal.com · 28/10/2025
TIL btop supports GPUs! github.com/aristocratos...
010
Antón @anton.galiglobal.com · 27/10/2025
Been testing GitHub Copilot’s new coding agent for a few days. Integration with Issues, PRs, Actions, Codespaces, and the mobile app is impressive. Still prefer CLI-based code agents, the review/feedback loop is much faster
010
Antón @anton.galiglobal.com · 23/10/2025
Set up my G Pro with gpro-led so the first letter of each of my kids’ names glows in their favorite color. They noticed instantly and loved it
020
Antón @anton.galiglobal.com · 22/10/2025
Neat trick: use watchexec to detect commits in your local repo and get notified when your code agent (running in a container) finishes its work.
020
Antón @anton.galiglobal.com · 22/10/2025
I've created my first Firefox plugin that lets me share the current tab to @frontpage.fyi with minimal effort (mobile and desktop). It's also the first time I use @atprotocol.dev and it's as cool as it seems. The repo is available in Github but also in @tangled.org: tangled.org/@galiglobal....
Plugin screenshot
2485
Antón @anton.galiglobal.com · 20/05/2025
What a special moment to speak at Current one last time about New Relic's object storage adoption for data streaming. Looking back at four years of performance improvements has been incredible, and I'm amazed by how much I've learned!
050
Antón @anton.galiglobal.com · 03/03/2025
It took me ages to finish Database Internals. Very tough read, I almost give up a couple of times. It covers way too much for a 300 pages book. Yet, it's great to have a book on the topic and I appreciate all those links to papers and references to OSS systems
020
Antón @anton.galiglobal.com · 28/02/2025
Understanding Custom Triggers In Apache Flink dev.to/ipolyzos/und...
000
Antón @anton.galiglobal.com · 27/02/2025
It's surprising Flink doesn't have support for Flatbuffers yet, it seems like a potential good candidate for great performance github.com/apache/flink...
110
Antón @anton.galiglobal.com · 27/02/2025
I was working on a serialisation formats benchmark this week and surprised how efficient Flatbuffers is. Trying to understand it, I found this article about Zero-Copy serialisation. I was already aware of Cap'n Proto but I didn't know Flatbuffers also use it. brunocalza.me/what-zero-co...
221
Antón @anton.galiglobal.com · 26/02/2025
AWS MSK and Confluent Cloud support? Check! In combination with the Java Agent Kafka monitoring, it's like peanut butter and jelly for your streaming stack. Battle-tested internally and ready to roll. More integrations incoming, and hopefully a little squirrel 🐿️ at some point
000
Antón @anton.galiglobal.com · 26/02/2025
I'm one of those annoying engineers always pushing to make our internal tooling available for customers. It has been a long journey but I'm very proud of New Relic Queues & Streams finally being released! docs.newrelic.com/docs/message...
110
Antón @anton.galiglobal.com · 17/02/2025
"We utilize the ‘island model’ for deploying our Flink jobs, where all dependencies for a given application reside within a single region" After 4 years using this approach, it's my default choice nowadays (but isolating at the AZ level instead of region). netflixtechblog.com/introducing-...
020
Antón @anton.galiglobal.com · 10/02/2025
Iceberg FileIO is very interesting to optimize Object Storage reads/writes. I'm working on similar optimizations but the edge cases are tricky. I wonder if someone has already used it without Iceberg for similar use cases www.tabular.io/blog/iceberg...
000
Antón @anton.galiglobal.com · 07/02/2025
Revolutionizing Failure Management in Apache Flink: Meet FLIP-304's Pluggable Failure Enrichers ❤️ www.confluent.io/blog/flip-30...
021
Antón @anton.galiglobal.com · 06/02/2025
The YCombinator Requests for Startups is pretty interesting. It seems that being a System Engineer is going to be cool again www.ycombinator.com/rfs
010
Antón @anton.galiglobal.com · 03/02/2025
LLMs are impressive knowledge tools, but experience is still king
010
Antón @anton.galiglobal.com · 14/01/2025
Oh, they fixed it the day after my comment! 🎉
010
Antón @anton.galiglobal.com · 10/01/2025
Built a tool using GraalVM for native compilation and Langchain4j with Ollama/Mistral to parse command outputs and generate readable summaries. When tests fail, it pipes the processed data back to NeoVim/Copilot to suggest fixes. And yes, I'm doing it because it's fun. No regrets.
000
Antón @anton.galiglobal.com · 07/01/2025
Your suspicion is correct, there's a small link to them on the bottom right of the page
110
Antón @anton.galiglobal.com · 06/01/2025
This weekend's project: adding native Vim diff support to CopilotChat.nvim. The plugin is fantastic, but I felt it was missing this feature. The ability to use Vim shortcuts for reviewing LLM output offers a real productivity improvement!
120
Antón @anton.galiglobal.com · 27/12/2024
I've successfully gotten the #ForSt State Backend working, both locally and with S3. It's a bit of a pain because it requires activation for DataStreams when using enableAsyncState(). The SST files are visible now. I'll next try to open them from a separate process (no Flink) as secondary instances
010
Antón @anton.galiglobal.com · 25/12/2024
I've been exploring Flink's new Disaggregated State Storage (FLIP-423). It designates object storage as the authoritative source for state, enabling direct querying. Local disks are no longer mandatory but can serve as a local cache. Early proof-of-concept results show promising performance:
150
Antón @anton.galiglobal.com · 17/12/2024
TableDescriptor is definitely my favourite way to create tables in Flink with the Table API because of the type safety and connectors options discoverability
010
Antón @anton.galiglobal.com · 17/12/2024
More knobs to turn! > When increasing the bandwidth for one of the service, say VPC, the available bandwidth for the other service, EBS, is reduced. IBC does not increase your instances burst bandwidth, packets per second, or IOPS capabilities. aws.amazon.com/about-aws/wh...
000
Antón @anton.galiglobal.com · 15/12/2024
Be careful. Slack workflows and lists are addictive!
000
Antón @anton.galiglobal.com · 11/12/2024
Apache Flink 🐿️ makes so easy to write tests and run them locally with the IDE or Maven/Gradle ❤️
110
Antón @anton.galiglobal.com · 11/12/2024
DuckDB's time_bucket function simplified hourly wind power aggregation by zone. The query effortlessly grouped and summed the data. The extensions model is a great idea and there are already plenty of contributions duckdb.org/docs/extensi...
DuckDB's time_bucket function simplified hourly wind power aggregation by zone from a JSONL file. The query effortlessly grouped and summed the data.
020
Antón @anton.galiglobal.com · 11/12/2024
Need to sum fields in a struct with varying numbers of elements? DuckDB's flexibility made it trivial!
110
Antón @anton.galiglobal.com · 11/12/2024
DuckDB made my last pet project a breeze! This is now a must-have in my data toolkit. Effortlessly explored JSON, CSV, and Parquet files, both locally and from object storage 👇
150
Antón @anton.galiglobal.com · 08/12/2024
I've just finished the book Algorithms to Live By. It's fun to see computer science applied to everyday problems. The last chapter about game theory was my favorite, I learned about the Vickrey auction and how it's designed to make honesty the best strategy
170
Antón @anton.galiglobal.com · 30/10/2024
There's a clear trend to migrate from Avro to Protobuf for streaming pipelines. This presentation by Uber was a great example. They showed several improvements they are contributing to Flink to make it possible: www.flink-forward.org/berlin-2024/...
296