Victoria Slocum
@victorialslocum
learning cool stuff, machine learning engineer at @weaviate_io ๐
Text-to-SQL is dead ๐ The next generation of querying is agentic - and itโs already here. This paper (arxiv.org/abs/2502.00032) introduces a type of agentic querying called ๐๐๐ป๐ฐ๐๐ถ๐ผ๐ป ๐๐ฎ๐น๐น๐ถ๐ป๐ด that uses an LLM to structure queries using predefined function calls inโฆ
We built a skincare app ๐ฑ๐ถ๐ณ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ ๐๐ต๐ฎ๐ป ๐ฎ๐ป๐๐๐ต๐ถ๐ป๐ด ๐ฒ๐น๐๐ฒ on the market. But it took a lot more than just slapping AI on a data pipeline. We just released ๐๐น๐ผ๐๐ฒ - and it's not your typical beauty app. It's what happens when you combine domain knowledgeโฆ
Think you need to code to build agentic AI systems? Not anymore. ๐ช๐ฒ ๐ท๐๐๐ ๐น๐ฎ๐๐ป๐ฐ๐ต๐ฒ๐ฑ ๐ฎ ๐ณ๐ฟ๐ฒ๐ฒ @n8n_io ๐๐ฒ๐บ๐ฝ๐น๐ฎ๐๐ฒ ๐๐ต๐ฎ๐ ๐ฝ๐ฟ๐ผ๐๐ฒ๐ ๐ถ๐. It's an agentic workflow that automatically sends you a weekly email summarizing the latest in AI research. But whatโฆ
90% of AI chat assistants suck. So why is Gloweโs so good? Because itโs fundamentally different from every other chatbot you've tried before. Here's the thing: ๐บ๐ผ๐๐ ๐๐ ๐ฎ๐๐๐ถ๐๐๐ฎ๐ป๐๐ ๐ฎ๐ฟ๐ฒ ๐ท๐๐๐ ๐ณ๐ฎ๐ป๐ฐ๐ ๐๐ฒ๐ฎ๐ฟ๐ฐ๐ต ๐ฏ๐ฎ๐ฟ๐. They match keywords, retrieveโฆ
Ouuf... What a vibe check at the Berlin Hackathon with @techeurope_ at @DeltaCampTeam ๐ฎโ๐จ 120 Hackers building with @weaviate_io @OpenAI @elevenlabsio and many more awesome companies ๐ฅ What a fantastic atmosphere with amazing people @philipvollet @victorialslocumโฆ
Bringing the right people together and create the energy Berlin needs to build @techeurope_ ๐ค Thanks @blwiertz for making this possible!
๐๐ฆ๐ฎ๐ช๐ฏ๐ช ๐๐ฎ๐ฃ๐ฆ๐ฅ๐ฅ๐ช๐ฏ๐จ๐ด ๐ค๐ค ๐๐ฆ๐ข๐ท๐ช๐ข๐ต๐ฆ We just published a new recipe notebook showing you how to use Google's Gemini embeddings with Weaviate. โข ๐ญ๐ฌ๐ฌ+ ๐น๐ฎ๐ป๐ด๐๐ฎ๐ด๐ฒ๐ ๐๐๐ฝ๐ฝ๐ผ๐ฟ๐๐ฒ๐ฑ: Build truly multilingual search experiences โข ๐ ๐ง๐๐โฆ
If youโre not already building with @n8n_io, then wtf are you doing? n8n.io/workflows/5817โฆ
We just killed AI subscriptions One wallet. Every tool. No subscriptions. Just sign in. Early Access starts NOW
No code Weaviate is here. And itโs part of the worldโs favorite automation platform: @n8n_io. Our new community node changes everything. Hereโs what it can do: 1. Insert documents into a Weaviate collection. 2. Return ranked documents based on queries from a Weaviateโฆ
What does it ๐ฎ๐ฐ๐๐๐ฎ๐น๐น๐ take to build a production-level AI-first app? Not another demo. Not a weekend hackathon project. A real app with real users solving real problems. We just launched Glowe (glowe.app), an AI-powered Korean skincare recommendationโฆ
Is this a beauty & life account now? No. But my colleagues built a skincare agent. It's called Glowe. And it's so good, it recommended a moisturizer I already use and like. How is Glowe so good? - They've analyzed almost 100,000 product reviews using @GoogleDeepMind'sโฆ
What happens when you build an AI-first app from scratch? Constella asked themselves this exact question two years ago - not how to add AI features to an existing app, but how to create a development stack that's ๐๐ฟ๐๐น๐ ๐๐-๐ป๐ฎ๐๐ถ๐๐ฒ. They wanted to create an "infiniteโฆ
Ever felt overwhelmed in finding your perfect skincare routine? We just launched ๐๐น๐ผ๐๐ฒ - an AI-powered Korean skincare app that actually understands what your skin needs. Try it out here: glowe.app/?utm_source=chโฆ Here's what makes Glowe special: ๐ฏโฆ
Most AI initiatives in healthcare are a HIPAA violations waiting to happen. The tension between innovation and risk has paralyzed the industry. Until now. Teams are either: 1. Scrambling to bolt on security to tools never designed for Protected Health Information (PHI). 2.โฆ
Looking for a cheaper, open source alternative to agentic RAG? Try multi-vector retrieval with Reason-ModernColBERT. Most search systems compare one big summary vector per document. Multi-vector retrieval is different - it keeps separate vectors for each word or phrase, thenโฆ
Today I'm celebrating 2 years at @weaviate_io! ๐ Here are 37 things I learned about retrieval so far. ๐leoniemonigatti.com/blog/what_i_leโฆ
Stop throwing better models at your RAG problems. Start optimizing what happens BEFORE retrieval instead. The secret isn't just in better embeddings or fancier models - it's in how you optimize queries before they hit your database. Here are 4 pre-retrieval techniques toโฆ
What if I told you that you could slash your multi-vector embedding costs by 70%? (and still maintain solid performance?) Multi-vector models like ColBERT and ColPali are incredible - they capture way more semantic information than single-vector models by preserving token-levelโฆ
This week was special! I had the honor of leading the @GoogleDeepMind Gemini session on the main stage during I/O Connect Berlin๐ฉ๐ช I could show everything our API has to offer, including our new Gemini 2.5 model family and advanced features such as text-to-speech, Live API, andโฆ
Thereโs just something about Google events that always gets me inspiredโฆ Had an absolutely AMAZING time at #GoogleIOConnect in Berlin yesterday! Really appreciate all that @googledevs is doing for us builders, between the Agent Development Kit to the Gemma models to theseโฆ