Anas Dorbani

PhD Student @dais_polymtl & @Mila_Quebec | Building Flock | Full details & code: | Ex. Research Intern @Oracle

[Anas Dorbani @anasdorbani]

🚀 We just released Flock v0.7.0, a @duckdb community extension that brings AI operators via LLMs, and RAG directly into SQL.

🧵 Here's what's new 👇

[DAIS 𓁈 @ Polytechnique Mtl @dais_polymtl]

Glad to see @lbugdb's performance improvements, and especially happy to see our recent work being adopted.

@lbugdb is adapting ideas from our recent SIGMOD 2026 paper to improve vectorized execution over factorized intermediates.

(1/2)

[Omar Sanseviero @osanseviero]

As part of the Gemma 4 release, we're launching Agent Skills: an Android app experience where you can import different skills and have Gemma 4 E2B reason and use the skills!

Running entirely in the phone, available in the Google PlayStore. Try it now!

[Tom Dörr @tom_doerr]

Automates poster generation from papers

https://t.co/wnfcurLyFN

[PhD_Genie @PhD_Genie]

When your paper gets published

[Anas Dorbani @anasdorbani]

@justbyte_ Books + Learning by doing 😁

[Mo Malekpour @momalekpour]

Happy to share that our recent work at @DAIS_PolyMTL, “SQLMorph: Query Mutation and Fine-Grained Metrics for Text-to-SQL Evaluation” has been accepted at @ICDEconf 2026, one of the top venues in data engineering. 🧵👇🏻

[Anas Dorbani @anasdorbani]

@priyaaa63 @duckdb The project is still a work in progress, but all supported functions are fully operational and tested. We are actively addressing any bugs across multiple platforms and look forward to receiving your feedback 😁.

[Anas Dorbani @anasdorbani]

🚀 We just released Flock v0.7.0, a @duckdb community extension that brings AI operators via LLMs, and RAG directly into SQL.

🧵 Here's what's new 👇

🤖 Anthropic (𝗖𝗹𝗮𝘂𝗱𝗲) 𝗣𝗿𝗼𝘃𝗶𝗱𝗲𝗿 𝗦𝘂𝗽𝗽𝗼𝗿𝘁

Flock now supports four LLM providers: OpenAI, Azure, Ollama, and Anthropic. Define a model once with CREATE MODEL, then swap providers or the model later (admin-side) without changing the SQL queries that use it.

🌐 𝗪𝗔𝗦𝗠 (𝗪𝗲𝗯𝗔𝘀𝘀𝗲𝗺𝗯𝗹𝘆) 𝗦𝘂𝗽𝗽𝗼𝗿𝘁

Flock now compiles and runs inside DuckDB-WASM.

📊 𝗟𝗟𝗠 𝗠𝗲𝘁𝗿𝗶𝗰𝘀 𝗧𝗿𝗮𝗰𝗸𝗶𝗻𝗴

End-to-end observability for your pipelines: token usage, latency, and call counts for all LLM invocations within a query.

🎤 Introduced 𝗔𝘂𝗱𝗶𝗼 𝗧𝗿𝗮𝗻𝘀𝗰𝗿𝗶𝗽t𝗶𝗼𝗻

Expanded multimodal support with audio transcription (and continued support for images).

If you want semantic and analytical processing in one place, Flock lets you do it all in SQL.

⭐ GitHub: https://t.co/R3surljdOJ
📖 Docs: https://t.co/oqTyAGM3Sf
📝 Paper: https://t.co/eu2g2Thaq4

We'd love your feedback and contributions!

#DuckDB #LLM #SQL #OpenSource #ArtificialIntelligence #RAG #DataEngineering #Anthropic #Claude #WASM #MachineLearning

[kai @watanabe_kai_]

job market so bad I'm actually pursuing my dreams

[JNS @_devJNS]

Vibe coders after realizing they'll still have to dance on TikTok to market their SaaS.

[Anas Dorbani @anasdorbani]

@thekaranpargaie The critical step AI misses is often the unstated context we instinctively provide.

[Anas Dorbani @anasdorbani]

How to bring AI directly into your SQL workflows? Explore open-source extensions that embed large language model (LLM) capabilities for RAG and semantic analysis. This approach lets data engineers and developers leverage AI in-database, simplifying advanced analytics within familiar SQL.

https://t.co/bsOFHzwMTh

[Anas Dorbani @anasdorbani]

Most data engineers are still wrestling with separate tools for AI and SQL. You're bleeding time and performance. The old way is dead. Here's why: https://t.co/uLn8ChaTZ8

[Anas Dorbani @anasdorbani]

Your 'modern' data stack is obsolete if it can't natively handle AI. Python notebooks are a temporary fix. True data engineering integrates AI directly into SQL. The future demands it.

[Anas Dorbani @anasdorbani]

SQL can feel like a chef trying to whip up a Michelin-star meal with only a spatula and a dream.

But with the right tools, even complex data transformations become a culinary masterpiece.

Don't let your data be a bland dish; spice it up with smart SQL!

[Anas Dorbani @anasdorbani]

@IAmAaronWill Wow, that's amazing!

[Anas Dorbani @anasdorbani]

@Yannick_Veys I completely agree with this!

[Anas Dorbani @anasdorbani]

@pcshipp They've already set the standard, and while others are trying to reach it, they're still surpassing them.

[anton @abacaj]

“Make the models cheap to use”
“Great, they all forgot how to code”
“Now 10x the price”

[Anas Dorbani @anasdorbani]

If you want semantic and analytical processing in one place, Flock lets you do it all in SQL.

⭐ GitHub: https://t.co/SD6kxW4vDm
📖 Docs: https://t.co/CSe0UwZwOQ

We'd love your feedback and contributions!

[Anas Dorbani @anasdorbani]

🚀 We just released Flock v0.7.0, a @duckdb community extension that brings AI operators via LLMs, and RAG directly into SQL.

🧵 Here's what's new 👇

[Anas Dorbani @anasdorbani]

🎤 Introduced 𝗔𝘂𝗱𝗶𝗼 𝗧𝗿𝗮𝗻𝘀𝗰𝗿𝗶𝗽t𝗶𝗼𝗻 Expanded multimodal support with audio transcription (and continued support for images).