Ojasvi Agarwal’s Post

Gone are the days when you would Google important things/life advice like "How to transition from SWE to ML Engineer" without running into overly optimized trite SEO Optimized articles. I find good and important information on Reddit/Twitter, but then, I feel the information is scattered and unorganized again. Introducing TextData, a new tool made by Kevin Ros on which I am contributing these days. It allows you to save your notes, bookmark those hundreds of tabs scattered across your desktop, and, most importantly, build public communities and creators (or organizers? haha) can come up with curated content compilations. The vision is to create a community marketplace where creators could come up and develop roadmaps, and organize gated knowledge/personal notes that eventually would benefit members. Do give it a spin! #BuildInPublic! What are your thoughts on the platform?

Introducing TextData! TextData is a social platform for collaboratively creating, sharing, and learning from wiki-style communities. We've all encountered the problem of learning about new topics online - it usually requires countless hours of searching, browsing, and sifting through low-quality websites, SEO-optimized spam, and confusing content.  TextData hopes to solve this, and to reward those that spend their time curating and organizing online resources. With TextData, you can create a community for a particular topic, task, or goal, and organize the world's web resources there. Then, if you want, you can share this community with the world, letting anyone with a TextData account benefit from your knowledge. You can also follow public communities to stay up-to-date with the organized content. And you can search, visualize, and see recommendations from your followed communities, ensuring that you can find the content when needed. For more information, check out our website at https://textdata.org/about And once you make an account, be sure to start with the TextData onboarding community: https://lnkd.in/gJ_gvS7y

Ever felt overwhelmed by the hundreds of bookmarks cluttering your browser? Ever struggled to find that one bookmark you need? Are you tired of trying to keep everything organized? Throughout the years I tried many bookmarking solutions and never found one I loved. So I decided to build one... Introducing Memxi Bookmarks! With Memxi, saving and retrieving websites is effortless. Imagine asking, "What's the marketing automation tool I bookmarked?" or "Which site had those AI voice cloning tools?" and getting your answers instantly. Memxi understands the content of your bookmarks, making everything you save easily accessible through natural language queries. Or even ask it questions about the actual content such as: "What was the price of the cell phone plan I saved?" But there's more: - Bookmark YouTube videos, and Memxi will transcribe them, letting you find content by what's said in the video. - Overwhelmed by a lengthy article? Get a concise summary with just one click. - Curious about something on a page? Ask Memxi, no bookmark needed. - If your bookmarks don't hold the answer, seamlessly query Chat-GPT directly from one unified search interface. And most importantly - your data is ALWAYS your data. Importing and exporting bookmarks is straightforward with Memxi, ensuring you can always take your data wherever you go. Want to also save bookmarks on mobile? Hit me up for early access to the mobile app version. Memxi is free, and the first thousand users to install get free access to pro features, including Chat-GPT 4, and more. All I ask is to send me feedback when you have a chance. Would love to get your help spreading the word and any feedback on the product. Thanks and happy bookmarking! https://lnkd.in/g-8yybn6

Hello everyone, I'm thrilled to announce the release of my latest project on Social Network Ads classification, now live on GitHub! This project is a culmination of extensive research and application of advanced machine learning algorithms to achieve superior accuracy and performance. Key Features and Highlights: Outstanding Accuracy: Achieves an impressive 95% accuracy in classifying social network advertisements. Diverse Algorithms: Implements a variety of machine learning techniques, including Linear Regression, Polynomial Regression, Support Vector Machine (SVM), Decision Trees (DT), Random Forest, and XGBoost. This diverse approach ensures robust and reliable predictions. Optimization Techniques: Utilizes Grid Search and Random Search for hyperparameter tuning, optimizing model performance to its highest potential. Rigorous Data Preprocessing: Comprehensive preprocessing and data cleaning steps are employed to eliminate noise and ensure high-quality inputs, leading to more precise predictions. Deployment Ready: The model is fully prepared for deployment, making it seamless to integrate into applications and operational environments. Wide Application: This tool is invaluable for marketers, advertisers, and data analysts, providing data-driven insights to enhance the effectiveness of social network ad campaigns. This project empowers businesses with the capability to gain accurate insights into their ad performance, facilitating informed decision-making and optimizing advertising strategies. Explore the project, review the code, and see it in action by visiting the GitHub repository: https://lnkd.in/dWBghw9h Your feedback and thoughts on this project are highly appreciated. Your insights will be instrumental as I continue to refine and enhance this tool. #MachineLearning #DeepLearning #AI #SocialNetworkAds #DataScience #RegressionModels #SupportVectorMachine #DecisionTrees #RandomForest #XGBoost #ModelOptimization #GridSearch #RandomSearch #DataPreprocessing #ArtificialIntelligence #GithubProjects Eager to hear your feedback!

Google News, is a hub for real-time updates with curated news, personalized feeds, and trending topics. It aggregates news to help users stay informed about the latest stories and events. Google News attracts over 370 million monthly visitors, making it a globally influential platform for scraping insights, tracking trends, and extracting valuable data efficiently. In this blog, we’ll learn how to scrape Google News using Crawlbase’s Smart Proxy. By using JavaScript and the Smart Proxy, we collect datasets like article titles, URLs, publication dates, summaries, author info, and much more. Jump on this interesting read... 👇 https://lnkd.in/d33sB3UV #googlenews #WebScraping #DataScraping #GoogleNews #ScrapingEthics #DataMining #WebCrawling #DataExtraction #ScrapingTools #ScrapingTutorial #ScrapingBestPractices

The final strike to keyword search: 1 keyword doesn't always mean just 1 thing. There are many meanings. Take the word "process": + A series of steps taken to achieve a goal + A natural development involving change  + A system for accomplishing something + An operation executed by a computer + The act of dealing with paperwork + A biological reaction All of them are logical meanings of process. But it needs a short description to frame it. Only now you can imagine which process is meant. It's context. Right? To solve the keyword problem find a search engine: + Takes written text instead of keywords + Trained on multiple domains instead of only patents + Use the given context instead of extracting keywords from it Just don't get stuck with keywords. Ok? But there's another problem. Language evolves over time. And new expressions emerge. The keyword "App" didn't exist until 2008. But the concept was already known. Unstuck your search. Use context. ~~~ ✍️ Curious — do you agree? ♻️ Reshare if this was helpful to you.

The problem with Keywords The problem is... The keyword doesn't always mean just 1 thing. There are many meanings. Take the word "process": + A series of steps taken to achieve a goal + A natural development involving change + A system for accomplishing something + An operation executed by a computer + The act of dealing with paperwork + A biological reaction All of them are logical meanings of process. But it needs a short description to frame it. Only now you can imagine which process is meant. It's context. Right? To solve the problem find a search engine: + Takes written text instead of keywords + Trained on multiple domains instead of only patents + Use the given context instead of extracting keywords from it Just don't get stuck with keywords. Ok? But there's another problem. Language evolves over time. And new expressions emerge. The keyword "App" didn't exist until 2008. But the concept was already known. Unstuck your search. Use context. --- This is how you de-risk your strategy in 2023.

New free research tool for a tonne of blog ideas! Just enter a few seed terms and it will generate & group related keywords to form topics tailored to your niche. For the last ~6 weeks I have been playing with building out my own SAAS. Describe your business and it will guide you through a seed research process to then generate and output topics based on your seed terms. Leveraging SERP data, it groups keywords into topics and pulls in information like topic search volume, intent, topic competitiveness and trend data. It's taking longer than expected to refine the flow, but what a learning experience! So I thought I would release a more basic version for now. A version that generates keywords in a similar way, but rather than leveraging SERP data, it leverages semantic grouping processes. Basically, similar words/phrases get grouped together into a topic. It looks at the words, rather than any actual correlation of any ranking data between the keywords. Gets you most of the way there, but there is some manual work at the end to clean things up and extend the data. It allows me to keep the costs down and release a preliminary tool that still offers a way to find, filter & produce topics for free. Enter up to 3 seed terms, ideally closely related as that will work best, and then it will run through a generation process. Main tool will come, eventually, but for now - link in comments for the free blog topic idea generator. HOT TIP: If you want to game it a bit to spit out more topics for a single seed, then look at rotating word ordering & plurals. 'bed sheets' generates 85 topics, but then 'bed sheets, bed sheet, sheet bed' generated double that!

🤖 Introducing My Health Bot: Your Personal Doctor! 🚀 I’m thrilled to share my latest creation—a cutting-edge health bot that combines the power of LLM, Google Gemini Pro, and other advanced technologies. Let’s dive into what makes this system truly remarkable: Medical Report Analysis:📊 Users can effortlessly upload their medical reports. 🧐 If you’ve ever struggled to interpret a report, worry not! Just provide it to my model. 🤖 The bot expertly analyzes the report, determining whether it indicates a positive or negative outcome. Symptom-Based Assistance:🤧 Imagine those days when you’re battling a cold and feeling lost. We’ve all been there! 📝 Simply input your symptoms, and our health bot responds with personalized advice. 💊 You’ll receive recommended medications, precautions, and effective remedies for relief. Behind the Scenes:🌐 Our tech stack includes LLM (Google Gemini Pro), Langchain, Python, and Docker. 🗄️ Efficient file storage using Vector DB ensures seamless operations. 👉 Give It a Try! Whether you’re curious or genuinely seeking health guidance, interact with our bot. I’m confident it’ll exceed your expectations Need to fine tune the prompts and storing the embeddings ! 🌟 Here's Git Hub Repo = https://lnkd.in/d-rBRMic Feel free to add more things appreciate Ur suggestions.

I know what you were all thinking..."It's been a while since Francis shared an opensource AI powered marketing solution." Well worry not... 🚀 I’m excited to share an #opensource solution to help you personalize your marketing with Reinforcement Learning. It’s now available on #GitHub and it is called ✨ Optimus ✨. 💡 Imagine offering customized product recommendations or promotions in a data-driven way – that's personalized #marketing. Yet, marketers might find it hard to find the right #AI tools to do this. Hence, they miss an opportunity to further optimize their marketing efforts. 🤔 So we asked ourselves: Can we build an easy-to-deploy AI solution for this? ✅ That’s why we developed the open-source #Python library Optimus. It helps marketers to personalize their applications in real time (for dynamic landing pages, prices, etc.) with AI. Optimus is easy to deploy and it uses #ReinforcementLearning to constantly learn and improve. Learn more about it in our #AvailableOnGitHub series and in the documentation on GitHub below. 🔗 https://lnkd.in/ef6dRa7M #Google #MarTech #DataScience #GoogleCloud