RipeSeed

Our team recently worked with Antonio, one of our clients, to launch an MVP of an AI product. It's always a pleasure working with clients like him, who are super passionate and so kind with their words. Listen to what he has to say. #startups #mvp #productlaunch #clientreviews

A little behind-the-scenes of our incredible team in action! Driven by creativity, collaboration, and commitment. #team #officeculture #collaboration #engineering #developers

Wonder how AI can improve education or power up your Ed-Tech startup? AI is transforming the education system. With personalized learning and innovative teaching methods, AI is revolutionizing how we approach modern learning and opening up new possibilities for the future. The shift in the education landscape with AI has introduced effective solutions such as course planning, tutoring assistants, and personalized learning plans, simplifying day-to-day tasks. The ed-tech industry is now an adaptive and modern landscape. #education #edtech #ai #llms #generativeai #chatgpt #openai

Book your meetings with the AI Assistant! We’ve added a new feature in Ask RipeSeed. Now, you can book appointments by chatting with an assistant! We used OpenAI's function calling to make this possible. Check out the repository link below, and don’t forget to give it a star if you like it! Link: https://lnkd.in/dYhfEFw8 #chatgpt #ai #llms #generativeai #openai

When you allow users to interact with your product across a web based interface and a browser plugin, you can get into the problem of managing their user sessions across the apps. To make the user experience as seamless as possible across both interfaces, you need to maintain the sessions across both products consistently. Read this article below where Nushirvan explains how he solved this problem while working with one of our customers. https://lnkd.in/dsy4hZD9 #development #chromeextension #browserplugin #crossplatform #startup

Ask RipeSeed is now Open Source 🌍✨ https://lnkd.in/d2SPnn-x A few weeks ago we announced the launch of Ask RipeSeed and everyone loved that. Today we are open-sourcing everything. Here's everything you can find in the repo: 💬 Complete NextJS codebase for the Chat UI 🔗 Integration with Pinecone ⁍ Markdown support in Chat 📖 Example of structured knowledge base for agencies 📌 Support to attach your own document 🌗 Light / Dark mode (obviously) 🐍 Python notebook with optimal chucking strategy We spent around 1.5 months building Ask RipeSeed, and then another 1.5 months revamping the UI (twice), improving the response time (made it 3x faster), and just obsessing over some final details. The aim was to launch an AI assistant that our customers could simply interact with and know if we could help them with their ideas. They could just ask questions or give their requirements and the AI assistant will respond with relevant portfolios (if any) But then we thought, wouldn't it be amazing if we could just open source it? Considering how the whole development community is so much reliant on all these open-source frameworks and tools, it'd be amazing to contribute something from our side as well. Feel free to break it apart, and build your own AI assistants (and if you like it, don't forget to leave us a star ⭐️ on Github) Honorable Mentions? You can see them as contributors on Github 👀 #opensource #ai #llm #openai #chatgpt #langchain

Open Source LLMs are everywhere, and we all love them. We can maximize their power by fine-tuning them on our own data, hence getting results according to our own business needs. Read this blog where our Software Engineer, Muhammad Noraiz, dives deep into how we can fine-tune any Open Source LLM. https://lnkd.in/d2J8P3uv #llm #opensource #llama #gemma #mistral

We postponed our release three times before finally launching Ask RipeSeed. When we kicked off Ask RipeSeed project we thought it would be a breeze. The prototype was built in just a week, and we felt confident about our progress. But as we started working on the design, it became clear that we wanted this product to be more than just functional—it needed to look and feel great too. We set a timeline for launch and began refining the design with multiple rounds of revisions. By the time we got back to development, we were already behind schedule. Despite the best efforts of our dev team, the final product didn’t meet the quality standards we’re known for. That was the first time we hit the brakes on the release. The mistake? We underestimated the project simply because it was internal. Determined to turn things around, the dev team went back to the drawing board. This time, they focused on improving app performance and enhancing the quality of LLM responses. We reworked our knowledge base, experimented with new chunking algorithms, and finally saw significant improvements. We were ready to launch again, but at the last minute, we discovered bugs and realized we could push the LLM response times even further. So, we pulled back once more. They say the third time's the charm. This time, the team worked tirelessly to fix bugs, implement response caching, add markdown support, and introduce LLM response streaming, all while enhancing the user experience. Finally, we had a product we were proud of, and we launched with confidence. Quality takes time—and demands focus and attention. We got lazy because it was an internal project, but this experience reminded us that every project deserves the same level of dedication, regardless of its scope. #LessonsLearned #QualityMatters #Teamwork #ProductDevelopment #askripeseed #ripeseed

Yesterday we launched Ask-RipeSeed. And to be honest, it should’ve launched like two weeks ago if it weren’t for some annoying technical issues holding us back 🤷♂️ When getting the response from GPT & Langchain, whenever the response size grew larger, we had to wait a few seconds before seeing the final message. The problem was rather straightforward and it was the same reason why all these chatbots stream the response instead of just instantly showing it. LLMs don’t generate the full message all at once. They stream one word at a time. Therefore, like ChatGPT, we could also stream the response and show that illusion of faster response. I took on the task and, after quick research, found that you can get a response in the form of a stream from OpenAI’s chat model, and that’s what I did. The next challenge was not only to render the message but also to save it asynchronously while keeping the sequence of chunks in local storage. So I simply put the chunks in a queue and called the saving method asynchronously, ensuring that, regardless of the sequence in which they run, the chunks would be saved sequentially at the end of the process. This approach avoided using await and prevented the rendering of messages from slowing down. The result? Reduced latency and streaming of messages while saving the chunks along the way, so if a response is interrupted, the user doesn’t lose the entire message 🔥 And all those days and nights paid off and Ask-RipSeed finally got launched. So happy to see something you worked on tirelessly go live and see people liking and appreciating it ✌️ Ask: https://ask.ripeseed.io/

We are launching Ask RipeSeed 🚀💬🎉 After working on so many AI Assistants for so many customers, we thought it was about time we built one for ourselves. "But why? What purpose does it serve?" - Got an idea? Just ask what you are looking for and it'll tell you if we can do it or not - Want to know about us? Ask and it'll bring up relevant information - Want to know if we have worked on any particular technology? Just ask and it'll bring up relevant portfolios We have been working on it for quite some time, but being an internal product, it was always deprioritized due to the flow of work from our customers. Because that's always the priority (and a good problem to have) For the last couple of weeks, we have been planning to launch, but then almost every time, we thought we could improve it a bit more. Sometimes it was UI, sometimes it was the response time, and sometimes handling some edge cases. But today, finally, we are officially launching it. It's still not perfect and we have some plans to improve it even more, but we are excited to have this in our portfolio. Go check it out 👉 https://ask.ripeseed.io Hats off to our rockstar engineering team that worked on it → Afaq Ahmad, Osama Riaz, Ahmed Shahram, Usman Ahmad & Usama Javed #ai #generativeai #productlaunch #llms #rag #chatbot