Freddy Guime
Freddy Guime
@fguime.bsky.social
Distinguished Principal Engineer @expediagroup.bsky.social . Previously of the financial industry. Co-host of @javapubhouse, podcast for the Java dev. Co-Author of Java 7 Recipes
360 followers205 following38 posts
AI isn’t just for chatbots. 🚫🤖 In my latest post, I share how LLMs can actually simplify real work—from classifying requests to summarizing reports. www.linkedin.com/pulse/unlear...
Ok question @tychobrahe.bsky.social. If I were to start reading warhammer novels where to start?
Just posted a new article on what does it mean to be an engineer in the era of LLM/AI #ai #leadership #llm #developer www.linkedin.com/pulse/throug...
Freddy Guime reposted
#offheap 91! We dive into OpenAI’s $3B Windsurf deal, the launch of the CVE Foundation, and big OSS updates: Spring AI 1.0.0, Jakarta EE 11 and major license shifts from Redis, Hibernate & OpenRewrite. Plus, RTO trends & more! 🎙️ #Java #OSS #AI #DevNews www.pubhouse.net/podcast/offh...
Freddy Guime reposted
🎙️ New Stackd Podcast Ep. 77 — “I Haven’t Been Fired Yet” w/ @starbuxman! We talk AI debugging, Java 24, Spring AI, WebAssembly, UK’s crime prediction tool, clean energy, dev tools & more! ☕💻 www.pubhouse.net/podcast/stac... #Java
@danvega.dev how about a tutorial for #springai and #llamaguard3? That would be amazing! (Hint hint 😉)
So my weekend project (ok multiple weekends) was to create an email classifier that help me organize more than 100k emails on my inbox. It's open sourced, use LLM for email classification and runs in your computer (no need to ship emails outside to anyone). Give it a spin! github.com/windust/mail...
I will proudly die on this hill. Steak fries are the best fries🍟. www.nytimes.com/2025/03/31/d...
I wonder what @smarks.bsky.social thinks? (am I a heretic for even thinking it? 😂)
Freddy Guime reposted
💘 Open source is looking for love! In our latest #OffHeap episode, @apache.org #eclipsefoundation, and @commonhaus.org battle it out to win over a new open-source project. Who will be the perfect match? 🌹✨ www.pubhouse.net/podcast/offh...
Freddy Guime reposted
Attend @fguime.bsky.social's session at Arc of AI Conference, and learn how to build an email filter and classifier that runs locally using Java and Ollama ✉️ www.arcofai.com/speaker/1d91... Use code JOIN-FREDDY-50OFF for a discount 📅March 31-April 3 🎟️ arcofai.com
Freddy Guime reposted
Freddy Guime reposted
⭐⭐⭐⭐⭐ "When your systems are designed to be used by normal engineers, all that excess brilliance they have can get poured into the product itself, instead of wasting it on navigating the system."
The IEEE republished a lightly edited version of my recent essay, "In Praise of 'Normal' Engineers". 🙌 spectrum.ieee.org/10x-engineer The greatest engineering orgs in the world are not the most pedigreed or top heavy, but the ones where normal engineers can move the business forward, day by day.
Freddy Guime reposted
New JavaPubHouse Episode! (Spring AI and Ollama!). Learn how to run these fancy LLMs in your own computer, and best of all, interact with them using Java and Spring (it's a breeze!) www.javapubhouse.com/2025/02/epis...
let's crowdsource (maybe someone ran into this)... have a kotlin project running in WSL, when using gradle works fine... BUT if using maven, it breaks (by ways of not finding the JDK and for example saying can't find Superclass Object). @intellijidea.com uses gradle for build in Kotlin WSL,
tried forcing the additional params to specify the jdk home but neither seems to be taking so I'm at a loss... when I try to delegate build to Maven, it "builds" but then it fails on run (with a cryptic indexoutofbound) plus is slow. Is there something we could do for Kotlin, Maven projects, in WSL?
maybe a question for #springai folks @michaelminella.com .. I noticed that Spring AI + Ollama + Tools works well if doing non-blocking (chat), but when switching to stream it hangs. I did check the tips on and do have ollama 0.51.. bug? (docs.spring.io/spring-ai/re...)
When using #springai with @ollama.bsky.social the order of the System and User messages is important! Having it out-of-order mangles the prompt formation in the Ollama back-end and you get a bunch of hit stop token" pending=[<|start_header_id|>] stop=<|start_header_id|> FYI @danvega.dev :P
ok, maybe a question for @danvega.dev (or anyone that works with #springai). I am using it with Ollama, and while looking at the ollama server logs, it calls /api/chat for a OllamaChatModel.call, but I would rather call the /api/generate (as it's a one-shot generation)...how to do that?
You are up to date.