descriptions

ZanSara · ZanSara · commit 2e25e324dbe9 · 2025-09-08T11:51:02.000+01:00
diff --git a/content/about.md b/content/about.md
@@ -11,9 +11,9 @@ Earlier in my career I've been Lead AI Engineer for [Kwal](https://www.kwal.ai/)
 I've been NLP Engineer at [deepset](https://www.deepset.ai/), a German startup working on NLP [since "before it was cool"](https://www.deepset.ai/about), where I was the [main contributor](https://github.com/deepset-ai/haystack/graphs/contributors) of [Haystack](https://haystack.deepset.ai/), their open-source framework for building highly customizable,  production-ready NLP and LLM applications.
 I started my career at [CERN](https://home.cern/), where I had the privilege of driving one [major decision](/publications/tucpr03/) to migrate the graphical interface's software of the accelerator's control systems from Java to PyQt, and then of helping a client department [migrate](/publications/thpv014/) to this stack. I have also worked on other infrastructure and data pipelines, some of which resulted in [publication](/publications/thpv042/).
 
-Outside of work I have too many [pet projects](/projects) to follow up with than the free time I can dedicate to them. I love science fiction and space exploration, I enjoy challenging hikes in nature and learning languages, as much as  such process can be enjoyed.
+Outside of work I have more [pet projects](/projects) than free time to dedicate to them, just as everyone else; I love science fiction and space exploration, I enjoy hiking and learning languages, as much as such process can be enjoyed.
 
-I speak native Italian and fluent English, but I've also learned French during my time at CERN, I'm studying Hungarian  for family reasons, and Portuguese because I currently live there. I still can understand some Russian and I have a  very basic understanding of Chinese, both from my teenage and university years.
+I speak native Italian and fluent English, but I've also learned French during my time at CERN, I'm studying Hungarian for family reasons, and Portuguese because I currently live there. I still can understand some Russian and I have a very basic understanding of Chinese, both from my teenage and university years.
 
 You can find my latest CV [here](/me/sara_zanzottera_cv.pdf). Check out also my [projects](/projects), my [publications](/publications) and my [talks](/talks). If you prefer newsletters you can find my posts also on [Substack](https://zansara.substack.com/).
 
diff --git a/content/posts/2021-12-11-dotfiles.md b/content/posts/2021-12-11-dotfiles.md
@@ -1,5 +1,6 @@
 ---
 title: "My Dotfiles"
+description: What Linux developer would I be if I didn't also have my very own dotfiles repo?
 date: 2021-12-11
 author: "ZanSara"
 featuredImage: "/posts/2021-12-11-dotfiles/cover.png"
diff --git a/content/posts/2023-09-10-python-verbix-sdk.md b/content/posts/2023-09-10-python-verbix-sdk.md
@@ -1,5 +1,6 @@
 ---
 title: "An (unofficial) Python SDK for Verbix"
+description: If you need a Python SDK for a verb conjugator, try this one while it's still alive.
 date: 2023-09-10
 author: "ZanSara"
 featuredImage: "/posts/2023-09-10-python-verbix-sdk/cover.png"
diff --git a/content/posts/2023-10-10-haystack-series-intro.md b/content/posts/2023-10-10-haystack-series-intro.md
@@ -1,5 +1,6 @@
 ---
 title: "Haystack 2.0: What is it?"
+description: December is finally approaching, and with it the release of a Haystack 2.0.
 date: 2023-10-10
 author: "ZanSara"
 series: ["Haystack 2.0 Series"]
diff --git a/content/posts/2023-10-11-haystack-series-why.md b/content/posts/2023-10-11-haystack-series-why.md
@@ -1,5 +1,6 @@
 ---
 title: "Why rewriting Haystack?!"
+description: "Before even diving into what Haystack 2.0 is, how it was built, and how it works, let’s spend a few words about the whats and the whys."
 date: 2023-10-11
 author: "ZanSara"
 series: ["Haystack 2.0 Series"]
diff --git a/content/posts/2023-10-15-haystack-series-pipeline.md b/content/posts/2023-10-15-haystack-series-pipeline.md
@@ -1,5 +1,6 @@
 ---
 title: "Haystack's Pipeline - A Deep Dive"
+description: What are Haystack's pipelines and how do they work?
 date: 2023-10-15
 author: "ZanSara"
 series: ["Haystack 2.0 Series"]
diff --git a/content/posts/2023-10-26-haystack-series-canals.md b/content/posts/2023-10-26-haystack-series-canals.md
@@ -1,5 +1,6 @@
 ---
 title: "A New Approach to Haystack Pipelines"
+description: "Haystack 2.0 comes with a brand new pipeline concept. Let's discover it!"
 date: 2023-10-26
 author: "ZanSara"
 series: ["Haystack 2.0 Series"]
diff --git a/content/posts/2023-10-27-haystack-series-rag.md b/content/posts/2023-10-27-haystack-series-rag.md
@@ -1,12 +1,13 @@
 ---
 title: "RAG Pipelines from scratch"
+description: "Let's build a simple RAG Pipeline with Haystack 2.0 by just connecting three components: a Retriever, a PromptBuilder and a Generator."
 date: 2023-10-27
 author: "ZanSara"
 series: ["Haystack 2.0 Series"]
 featuredImage: "/posts/2023-10-27-haystack-series-rag/cover.png"
 ---
 
-*Last updated: 18/01/2023 - Read it on the [Haystack Blog](https://haystack.deepset.ai/blog/rag-pipelines-from-scratch).*
+*Last updated: 18/01/2024 - Read it on the [Haystack Blog](https://haystack.deepset.ai/blog/rag-pipelines-from-scratch).*
 
 ---
 
diff --git a/content/posts/2023-11-05-haystack-series-minimal-indexing.md b/content/posts/2023-11-05-haystack-series-minimal-indexing.md
@@ -1,12 +1,13 @@
 ---
 title: "Indexing data for RAG applications"
+description: RAG apps need data to work. Let's see how to pre-process our data to make our Haystack 2.0 RAG pipeline perform even better.
 date: 2023-11-05
 author: "ZanSara"
 series: ["Haystack 2.0 Series"]
 featuredImage: "/posts/2023-11-05-haystack-series-minimal-indexing/cover.png"
 ---
 
-*Last updated: 18/01/2023*
+*Last updated: 18/01/2024*
 
 ---
 
diff --git a/content/posts/2023-11-09-haystack-series-simple-web-rag.md b/content/posts/2023-11-09-haystack-series-simple-web-rag.md
@@ -1,12 +1,13 @@
 ---
 title: "The World of Web RAG"
+description: What if our RAG application could fetch data directly from the web, live? Let's build this pipeline with Haystack 2.0.
 date: 2023-11-09
 author: "ZanSara"
 series: ["Haystack 2.0 Series"]
 featuredImage: "/posts/2023-11-09-haystack-series-simple-web-rag/cover.jpeg"
 ---
 
-*Last updated: 18/01/2023*
+*Last updated: 18/01/2024*
 
 ---
 
diff --git a/content/posts/2024-01-06-raspberrypi-headless-bookworm-wifi-config.md b/content/posts/2024-01-06-raspberrypi-headless-bookworm-wifi-config.md
@@ -1,5 +1,6 @@
 ---
 title: "Headless WiFi setup on Raspberry Pi OS \"Bookworm\" without the Raspberry Pi Imager"
+description: Setting up a headless Pi used to be simpler. Is it still possible to do it without the RPi Imager?
 date: 2024-01-06
 author: "ZanSara"
 featuredImage: "/posts/2024-01-06-raspberrypi-headless-bookworm-wifi-config/cover.png"
diff --git a/content/posts/2024-02-20-is-rag-all-you-need-odsc-east-2024-teaser.md b/content/posts/2024-02-20-is-rag-all-you-need-odsc-east-2024-teaser.md
@@ -1,5 +1,6 @@
 ---
 title: "Is RAG all you need? A look at the limits of retrieval augmentation"
+description: This blogpost is a teaser for my upcoming talk at ODSC East 2024 in Boston, April 23-25.
 date: 2024-02-21
 author: "ZanSara"
 featuredImage: "/posts/2024-02-20-is-rag-all-you-need-odsc-east-2024-teaser/cover.jpeg"
@@ -13,10 +14,6 @@ aliases:
     speechifyLink="https://app.speechify.com/share/ca710e7b-5af7-4733-b3d8-c012758bd171"
 >}}
 
-*This blogpost is a teaser for [my upcoming talk](https://odsc.com/speakers/rag-the-bad-parts-and-the-good-building-a-deeper-understanding-of-this-hot-llm-paradigms-weaknesses-strengths-and-limitations/) at ODSC East 2024 in Boston, April 23-25. It is published on the ODSC blog [at this link](https://opendatascience.com/is-rag-all-you-need-a-look-at-the-limits-of-retrieval-augmentation/).*
-
----
-
 Retrieval Augmented Generation (RAG) is by far one of the most popular and effective techniques to bring LLMs to production. Introduced by a Meta [paper](https://arxiv.org/abs/2005.11401) in 2021, it since took off and evolved to become a field in itself, fueled by the immediate benefits that it provides: lowered risk of hallucinations, access to updated information, and so on. On top of this, RAG is relatively cheap to implement for the benefit it provides, especially when compared to costly techniques like LLM finetuning. This makes it a no-brainer for a lot of usecases, to the point that nowadays every production system that uses LLMs in production seems to be implemented as some form of RAG.
   
 ![](/posts/2024-02-20-is-rag-all-you-need-odsc-east-2024-teaser/rag_paper.png)
diff --git a/content/posts/2024-02-28-create-anki-cloze-cards-with-custom-gpt.md b/content/posts/2024-02-28-create-anki-cloze-cards-with-custom-gpt.md
@@ -1,5 +1,6 @@
 ---
 title: "ClozeGPT: Write Anki cloze cards with a custom GPT"
+description: Writing good Anki cards is a chore. Let's bring LLMs to the rescue.
 date: 2024-02-28
 author: "ZanSara"
 featuredImage: "/posts/2024-02-28-create-anki-cloze-cards-with-custom-gpt/cover.png"
diff --git a/content/posts/2024-04-14-eli5-llms.md b/content/posts/2024-04-14-eli5-llms.md
@@ -1,5 +1,6 @@
 ---
 title: "Explain me LLMs like I'm five: build a story to help anyone get the idea"
+description: Let's explore a high-level way to tell clearly what LLMs are good for to the average pedestrian and help them reason about it.
 date: 2024-04-14
 author: "ZanSara"
 featuredImage: "/posts/2024-04-14-eli5-llms/cover.png"
diff --git a/content/posts/2024-04-29-odsc-east-rag.md b/content/posts/2024-04-29-odsc-east-rag.md
@@ -1,5 +1,6 @@
 ---
 title: "RAG, the bad parts (and the good!)"
+description: A summary of my recent talk at ODSC East about RAG, just in case you haven't heard enough of it already.
 date: 2024-04-29
 author: "ZanSara"
 featuredImage: "/posts/2024-04-29-odsc-east-rag/cover.png"
diff --git a/content/posts/2024-05-06-teranoptia.md b/content/posts/2024-05-06-teranoptia.md
@@ -1,5 +1,6 @@
 ---
 title: "Generating creatures with Teranoptia"
+description: Having fun with fonts doesn’t always mean obsessing over kerning and ligatures. Sometimes, writing text is not even the point!
 date: 2024-05-06
 author: "ZanSara"
 featuredImage: "/posts/2024-05-06-teranoptia/cover.png"
diff --git a/content/posts/2024-06-10-the-agent-compass.md b/content/posts/2024-06-10-the-agent-compass.md
@@ -1,5 +1,6 @@
 ---
 title: "The Agent Compass"
+description: Agent means everything and nothing in today's GenAI landscape. Let's shed some light on this topic.
 date: 2024-06-10
 author: "ZanSara"
 featuredImage: "/posts/2024-06-10-the-agent-compass/cover.png"
diff --git a/content/posts/2024-09-05-building-voice-agents-with-open-source-tools-part-1.md b/content/posts/2024-09-05-building-voice-agents-with-open-source-tools-part-1.md
@@ -1,5 +1,6 @@
 ---
 title: "Building Reliable Voice Bots with Open Source Tools - Part 1"
+description: "A deep look at the main challenges of building performant and cost effective voice bots."
 date: 2024-09-20
 author: "ZanSara"
 featuredImage: "/posts/2024-09-05-building-voice-agents-with-open-source-tools/cover.png"
diff --git a/content/posts/2024-10-30-building-voice-agents-with-open-source-tools-part-2.md b/content/posts/2024-10-30-building-voice-agents-with-open-source-tools-part-2.md
@@ -1,5 +1,6 @@
 ---
 title: "Building Reliable Voice Bots with Open Source Tools - Part 2"
+description: "A practical guide on the best techniques to build performant and cost effective voice bots."
 date: 2024-10-30
 author: "ZanSara"
 featuredImage: "/posts/2024-09-05-building-voice-agents-with-open-source-tools/cover.png"
diff --git a/content/posts/2025-05-12-beyond-hype-reasoning-models.md b/content/posts/2025-05-12-beyond-hype-reasoning-models.md
@@ -1,5 +1,6 @@
 ---
 title: "Beyond the hype of reasoning models: debunking three common misunderstandings"
+description: "This is a teaser for my upcoming talk at ODSC East 2025, \"LLMs that Think: Demystifying Reasoning Models\". If you want to learn more, join the webinar!"
 date: 2025-05-12
 author: "ZanSara"
 featuredImage: "/posts/2025-05-12-beyond-hype-reasoning-models/cover.png"
diff --git a/content/posts/2025-05-16-llama-eu-ban.md b/content/posts/2025-05-16-llama-eu-ban.md
@@ -1,5 +1,6 @@
 ---
 title: "Using Llama Models in the EU"
+description: The ban's terms are surprisingly not well known among users of these popular "open-source" LLMs.
 date: 2025-05-16
 author: "ZanSara"
 featuredImage: "/posts/2025-05-16-llama-eu-ban/cover.png"
diff --git a/content/posts/2025-05-21-vibecoding.md b/content/posts/2025-05-21-vibecoding.md
@@ -1,5 +1,6 @@
 ---
 title: "A simple vibecoding exercise"
+description: "Can GenAI help you finish your side-projects?"
 date: 2025-05-21
 author: "ZanSara"
 featuredImage: "/posts/2025-05-21-vibecoding/cover.png"
diff --git a/content/posts/2025-06-02-can-you-really-interrupt-an-llm.md b/content/posts/2025-06-02-can-you-really-interrupt-an-llm.md
@@ -1,5 +1,6 @@
 ---
 title: "Can you really interrupt an LLM?"
+description: "You never see that in the demos... why?"
 date: 2025-06-02
 author: "ZanSara"
 featuredImage: "/posts/2025-06-02-can-you-really-interrupt-an-llm/cover.png"
diff --git a/content/posts/2025-06-xx-real-life-subtitles.md b/content/posts/2025-06-xx-real-life-subtitles.md
diff --git a/content/posts/2025-09-10-playing-guess-who-with-an-llm.md b/content/posts/2025-09-10-playing-guess-who-with-an-llm.md
@@ -0,0 +1,9 @@
+---
+title: "Playing Guess Who with an LLM"
+description: "I expected a different kind of fun."
+date: 2025-09-10
+author: "ZanSara"
+featuredImage: "/posts/2025-09-10-playing-guess-who-with-an-llm/cover.png"
+draft: true
+---
+
diff --git a/content/projects/guess-who.md b/content/projects/guess-who.md
@@ -0,0 +1,15 @@
+---
+title: "Guess Who"
+description: Play "Guess Who" with your favorite LLMs!
+date: 2025-09-01
+author: "ZanSara"
+featuredImage: "/projects/guess-who.png"
+---
+
+LLMs nowadays are headed for superintelligence, they say. They should make good tabletop game partners then, I guess?
+
+In this repo you can find a simple implementation of a Guess Who game. If you're not familiar with the rules, here they are: each player receives a character and has to guess the opponent's character name by asking one yes/no question at a time. For example: "Is your character blonde?" or "Does your character have glasses?". Whoever guesses the opponent's character first wins. 
+
+A very simple game that should be within the understanding of any multimodal language model, right? [Try for yourself](https://zansara.dev/guess-who/) to find out.
+
+You can also find the game on [GitHub](https://github.com/ZanSara/guess-who). Contributions of all kinds are welcome 🙇
diff --git a/static/posts/2025-09-10-playing-guess-who-with-an-llm/cover.png b/static/posts/2025-09-10-playing-guess-who-with-an-llm/cover.png
diff --git a/static/projects/guess-who.png b/static/projects/guess-who.png
diff --git a/themes/hugo-coder/layouts/_default/single.html b/themes/hugo-coder/layouts/_default/single.html
@@ -12,25 +12,25 @@ <h1 class="title">
             </a>
           </h1>
         </div>
-        <div class="post-meta">
+        <!--div class="post-meta">
           {{ if .Date }}
           <div class="date">
             <span class="posted-on">
-              <i class="fa fa-calendar" aria-hidden="true"></i>
+              <!--i class="fa fa-calendar" aria-hidden="true"></i->
               <time datetime="{{ .Date.Format "2006-01-02T15:04:05Z07:00" }}">
                 {{ .Date | time.Format (.Site.Params.dateFormat | default "January 2, 2006" ) }}
               </time>
             </span>
             <!--span class="reading-time">
               <i class="fa fa-clock-o" aria-hidden="true"></i>
               {{ i18n "reading_time" .ReadingTime }}
-            </span-->
+            </span->
           </div>
           {{ end }}
           {{ with .GetTerms "authors" }}{{ partial "taxonomy/authors.html" . }}{{ end }}
           {{ with .GetTerms "categories" }}{{ partial "taxonomy/categories.html" . }}{{ end }}
           {{ with .GetTerms "tags" }}{{ partial "taxonomy/tags.html" . }}{{ end }}
-        </div>
+        </div-->
       </header>
 
       <div class="post-content">
diff --git a/themes/hugo-coder/layouts/posts/single.html b/themes/hugo-coder/layouts/posts/single.html
@@ -11,13 +11,16 @@ <h1 class="title">
               {{ .Title }}
             </a>
           </h1>
+          <p style="margin-bottom:0; color:#777;">
+            {{ .Description }}
+          </p>
         </div>
         <div class="post-meta">
-          <div class="date">
+          <div class="date" style="font-style: italic; margin-bottom: 2rem; line-height: 0.8; font-size: medium; color:#777;">
             <span class="posted-on">
-              <i class="fa fa-calendar" aria-hidden="true"></i>
+              <!--i class="fa fa-calendar" aria-hidden="true"></i-->
               <time datetime="{{ .Date.Format "2006-01-02T15:04:05Z07:00" }}">
-                {{ .Date | time.Format (.Site.Params.dateFormat | default "January 2, 2006" ) }}
+                by <a href="/">Sara Zan</a>, {{ .Date | time.Format (.Site.Params.dateFormat | default "January 2, 2006" ) }}
               </time>
             </span>
             <!--span class="reading-time">
diff --git a/themes/hugo-coder/layouts/shortcodes/audio.html b/themes/hugo-coder/layouts/shortcodes/audio.html
@@ -1,4 +1,4 @@
-<div style="border: 1px solid gray; border-radius: 10px; padding: 5px 15px; font-family: sans; font-style: italic;">
+<div style="padding: 5px 15px; font-family: sans; font-style: italic;">
   <p style="margin:5px 0 10px 0;"><a href="{{ .Params.audioFile }}">Listen to this article</a> or <a href="{{ .Params.speechifyLink }}?utm_campaign=partners&utm_content=rewardful&via=zansaradev-blog" target="_blank">read it in the app</a>:</p>
   <audio controls style="width:100%;">
     <source src="{{ .Params.audioFile }}" type="audio/mpeg">
