Sameh Methnani

[Mahmoud Mabrouk @mmabrouk_]

New Tutorial: Building test sets from production traces for LLM eval in @agenta_ai

When your LLM app is in production, your users find edge cases you never thought of.

Here I show how to:

- Convert (batch) traces → test sets
- Create test sets in the playground

[Sameh Methnani @sam_methnani]

@nawaat هذا مثلا صار يوم الانتخابات
https://t.co/3hdi7Gf3Gk

[Karim Benabdallah @karim2k]

الله يرحمك ياسر 🕊

[Henda Fellah @Hendouch87]

Best moment of the protest today.
This generation was born into freedom...and they will never settle down ✊🏼
#Tunisia #Tunisie #تونس

[Karim Benabdallah @karim2k]

Credits photo @YassineGaidi

[Sameh Methnani @sam_methnani]

Pantheon 🤯

[Riham Alkousaa @RihamKousa]

Another level of German bureaucracy: Germany’s embassy in Tunis did not give a Tunisian Hip Hop artist a visa to come to Germany to have a live concert in Berlin which the embassy ITSELF publicized on it’s Facebook page last month! 😅
@AuswaertigesAmt @GermanEmbassyTN

[Mahmoud Mabrouk @mmabrouk_]

Llama 3 now available in @agenta_ai playground!

Yesterday, @AIatMeta announced its latest and most powerful open-source model, #llama3. Today it's in the playground of @agenta_ai

To quickly compare and evaluate Llama3 vs. 70 different models, head to @agenta_ai (links 👇🏼).

[Mahmoud Mabrouk @mmabrouk_]

Product Update in @agenta_ai: Improved Evaluation Workflow

We've spent the past month re-engineering our evaluation workflow for #LLM apps. Here's what's new:

[Agenta @agenta_ai]

📢 Feature announcement: Comparing LLMs and Prompts: Cost, Latency, Tokens

Creating a great #LLM app isn't just about making it good; it's also about finding the cheapest and quickest option. The best teams aim to balance quality and cost.

[Mahmoud Mabrouk @mmabrouk_]

We've just launched chat capabilities in @agenta_ai. If you are building #chat applications and struggling with errors, #hallucinations, and wrong answers. Read ahead:

With our open-source platform agenta, you can now:
⭐ Develop and test chat applications in the playground
⭐ Compare various prompts and configurations
⭐ Easily add conversations from the playground to the test set
⭐ Evaluate #chatbots using human feedback, featuring an interface for A/B testing, scoring, and annotation
⭐ Implement automatic evaluations to check output quality before production

We support all #LLM providers, and can handle both simple chat applications and more complex ones based on #RAG.

Check it out: https://t.co/hMljKPWAwJ
Repo: https://t.co/EWQDPl7WFg

#llmops #evaluation #llm #openai

[Agenta @agenta_ai]

We're thrilled to announce the launch of the Cloud-hosted version of Agenta.

A huge thank you to our incredible team for their hard work over the past few weeks, and to our amazing 49 contributors to the project!

Check it out: https://t.co/dtdA86t5rL

[Agenta @agenta_ai]

🚀 New Features in #agenta! 🎉 This week, we've shipped several improvements to agenta's evaluation capabilities. Dive in for the details! ⬇️

[Agenta @agenta_ai]

A big shoutout to @GeekflareHQ for featuring us in the "9 Best Open Source LLMOps Platforms to Develop AI Models"
https://t.co/QRCnXn2jNM

[Mahmoud Mabrouk @mmabrouk_]

🚀 Just published a tutorial on building an AI agent that notifies you on Discord about @HackerNews posts that might interest you! Used @agenta_ai for the #LLM app and @novuhq for the messaging .
#AI #Tutorial #HackerNewsBot
https://t.co/JHhwziNIqD

[Mahmoud Mabrouk @mmabrouk_]

🎉 Exciting update from #Agenta! We've introduced a new feature to build production-ready #LLM apps in minutes, directly from our UI.🖥

Try the live demo here: https://t.co/1sg5IBYisd

#llmops #opensource

[Agenta @agenta_ai]

🚀Starting a tutorial series on building production-ready LLM apps! With @agenta_ai, you can build and deploy your @langchain LLM app in less than 2 mins.
Stay tuned for more tutorials on evaluating your app and building RAG-based LLM apps !🛠️💡#LLM #Langchain

[Agenta @agenta_ai]

We've thrilled to unveil our new feature in https://t.co/PqLj4M9bB7: AI Critic 🤖. It's a tool designed to evaluate the answers given by LLM apps.

[Mahmoud Mabrouk @mmabrouk_]

🚀 @agenta_ai 's Exciting Update: The #Evaluation Feature is here! 🧪 Building a robust #LLM application requires ample experimentation with #prompts and configurations. Our new feature is here to ease the process.

[Mahmoud Mabrouk @mmabrouk_]

🚀 New Agenta Feature Alert! 🚀

We're thrilled to announce that Agenta now allows you to upload and modify your test files directly from our UI!