Pathway

Running a startup even as a CTO means doing sales, and sales means drowning in slides, lots of slides 💯📜 . Turns out a solution is just a few lines of Pathway code away, and we made one: https://lnkd.in/d-Zk5ans Now we have all presentations indexed and searchable, it was so helpful that we are releasing the code - let me know if you give it a try 🚀!

Now you can index slides from all the presentations you have on your drive! - Turn each slide into an image - Use GPT-4o to extract data from the images - Index the results - Query the data The team from Pathway build an open-source app that it doing exactly that. Here's the link: https://lnkd.in/eGAY_EYu Check it out and give them a star! Thanks Pathway for letting me play with the app and for supporting our community!

Now you can index slides from all the presentations you have on your drive! - Turn each slide into an image - Use GPT-4o to extract data from the images - Index the results - Query the data The team from Pathway build an open-source app that it doing exactly that. Here's the link: https://lnkd.in/eGAY_EYu Check it out and give them a star! Thanks Pathway for letting me play with the app and for supporting our community!

Latest addition to Pathway's App Templates: MultiModal RAG Pipeline (MM-RAG) for PDFs with Tables! 📄 Traditional RAG struggles with important details within tables which is addressed by the method used here. The responses are significantly superior, and this can be tweaked to support charts, images, or optimized with rerankers, adaptive RAG, etc. You can run the pipeline as a containerized solution via Docker or quickly test it out within Google Colab. The steps are here: https://lnkd.in/d4aYpsqi  Pathway's incremental data pipeline here: – Extracts tables as images during the parsing process.  – GPT-4o explains the content of the table in detail.  – Saves the table content with the document chunk into the index, making it easily searchable. User queries are sent to the LLM with the relevant context including parsed tables. Pathway's in-memory incremental indexing preprocesses and indexes your data in real time, always providing up-to-date answers even if a new PDF is added or deleted from your data store. #RAG #finance #GPT4o #Pathway #multimodal #AI #GoogleColab #jupyter

Building Dynamic RAG Apps is a breeze with LangChain + Pathway! 🚀 Check out the App Template by Szymon Dudycz that helps you get started with your dynamic RAG app in Google Colab using your own data in minutes: https://lnkd.in/dPZb8vM3 LangChain is a popular framework for working on RAG applications. But as changes occur in data sources, developers often struggle a lot. ETL pipelines become messy, and keeping up with these changes becomes a headache for developers. Pathway's integration with LangChain solves this problem by ensuring your applications always provide up-to-date knowledge. With Pathway's incremental updates: ✔️ Easily monitor several data sources for any data changes (insertions/deletions/changes) ✔️ Instantly sync your RAG apps ✔️ Avoid complex ETL adjustments from Day 1 Try this app template within Colab and streamline your RAG solutions for production. Pathway is also available natively as a vector store within the LangChain ecosystem. Check the link in the comments for details! #RAG #DynamicRAG #LLM #ETL #Pathway #LangChain #Realtime

A few weeks ago, I had the pleasure of being on the BNP Paribas stage at Vivatech! The conversation focused on the power of AI and real-time data, emphasizing the critical role of "always fresh" data to ensure AI and ML models stay up-to-date and accurate. If you are building real-time intelligent applications and data pipelines, and need to integrate AI/ML models, let's chat! Full video: https://lnkd.in/egiTTSQj #BNPPForTech #AI #RealTimeData #BNPParibas

Now you can do Kafka ETL in Python! Kudos to this explainer and easy-to-run explainer with Pathway by Olivier Ruas, PhD. 🚀 Imagine you’ve been hired by a fraud-detection company monitoring logs from servers in New York and Paris. The logs have different time zones: You need to unify these different time zones into a single format to maintain data integrity. This is where ETL comes in. Kafka is a popular tool to build ETL pipelines that many companies use. But, it’s mainly used by Java and Scala developers, making it tricky for data scientists and ML engineers Enters Kafka ETL with Python using Pathway: Pathway, a faster framework built on Rust, can be used via a Python interface. As a Python developer, you can build ETL pipelines over Kafka in pure Python without compromising on performance. ✔️ Extract (E): Extract data streams from Kafka using Pathway Kafka input connectors. ✔️ Transform (T): Convert times with varying time zones into unified timestamps using Pathway’s datetime module. ✔️ Load (L): Load the final data stream back into Kafka. The entire script is available as a Pathway App Template, which can be run via Docker in minutes. Read the full tutorial and access the GitHub repository here: https://lnkd.in/gYB-NZwj #Pathway #Kafka #ETL #Python #DataEngineering #OpenSource

Viva Technology was a blast! It was great to discuss entrepreneurship in a fireside chat on Stage One and an honor to present the incredible Albane Dersy from inbolt as the Female Founder Challenge winner! Huge congrats, Albane! Similarly as last year, all of this right before Elon Musk's talk 😉 The three days at Vivatech were a fantastic opportunity to reconnect with happy customers like Transdev, CMA CGM, La Poste Groupe, and many others! We had a chance to discovered amazing, delighted users of Pathway (big shout-out to Mazars!). It was also a wonderful opportunity to deep dive with Innovation Leaders who are passionate about implementing AI with Pathway. Exciting news to come - stay tuned! #AI #data #Womenintech

Join us this Friday for our upcoming online AMA on "Stream Data Processing: Today’s Landscape and Horizons," organized in collaboration with our friends at the Indian Institute of Technology, Bombay – Web and Coding Club, IIT Bombay, and the Institute Technical Council, IIT Bombay, along with leaders from India’s leading fintech – National Payments Corporation Of India (NPCI). 🗓 When: 24 May (this Friday) at 12:30 PM CET / 4 PM IST 💻 Where: Online, via the LinkedIn event below ✅ What to Expect: – Insights from global and Indian perspectives on the stream data processing landscape. – Discussions on practical challenges, innovations, and future trends – Engaging in a Q&A session with the panel. ✅ About the Speakers: – Jayprasad Hegde: Head of Data Science and AI Initiatives at NPCI, expert in AI/ML and large-scale data solutions. – Adrian Kosowski: Chief Product Officer at Pathway, completed PhD at 20, co-founded Spoj, 15+ years in research. – Megh Vidani: Data Engineering Specialist at NPCI, expert in scalable streaming pipelines and Big Data technologies. 🔗 Mark "Attend" on the LinkedIn event below, and add it to your calendar. See you there! #AI #DataEngineering #StreamProcessing #Pathway #NPCI #IITBombay

D-2 to Viva Technology, Europe's biggest startup and tech event! Come and meet us at our booth K28-006, Hall 1 next to the Pitch Studio to talk about the future of LLMs, RAG, streaming, developer tools, real-time ML, you name it 👩🏻💻 See you there! Bonus: Following Zuzanna's victory at the Female Founder Challenge last year, she will be back on Stage One this year on Thursday, May 23, right after the pitches of the 2024 Female Founder Challenge. 📅 May 22-25 📍Parc des expositions de la Porte de Versailles, Paris.