Over the last few months there's a been a huge trend in implementing RAG for technical documentation and offloading support. In theory this makes sense to free up developers on rotation, but it's not as "simple" as you think. You need to have some sort of analytics behind your bot, decoupling queries from bot responses and synthesizing this information to continue extracting value from the voice of the customer. #voiceofcustomer #community #devrel #developeradvocate #chatbot https://lnkd.in/g4vv6WZS
Siftree’s Post
More Relevant Posts
-
Task 3: Real-Time Chat Application (Trixie Bot) Completed a real-time chat application as part of my journey with Pinnacle Labs, built using Flask. This project allowed me to dive deep into web development, creating a dynamic and interactive platform for real-time communication. I thoroughly enjoyed tackling the complexities of real-time data handling and enhancing the user experience. Looking forward to applying these skills to even more exciting projects! #RealTimeChat #FlaskFramework #WebDevelopment #PinnacleLabs #TechProjects #PythonDeveloper #SoftwareSolutions #CodingJourney #ChatApplication
To view or add a comment, sign in
-
The Chat with PDF application is a tool that allows users to upload PDF documents and interact with them by asking questions. Deployed application - https://lnkd.in/gj6ex_cS Github repo - https://lnkd.in/gkA-Z_ga #chatbot #genai
To view or add a comment, sign in
-
I've been working on writing a spec for in-app chat along with building a proof of concept for Code.Sydney's Marketplace App for Good. Spent some time thinking about: - How I want to query my data, - Where I want to put indexes on my tables, - Using ordered IDs for my Message IDs, - Polling for updates vs Websocket events (Pull vs Push communication) - State management in React using reducers and other alternatives. Also to be added in and not currently covered: - Using a last_message_id field on the threads table which would allow me to check a single row instead of scanning through the messages table. Not finished. Very on brand for me, hehehe. But also not bad for a spec considering that I've never been coached on writing specs. I just pieced together whatever I could find on the internet.
MA6 - In-App Chat · Issue #7 · codesydney/marketplace-app-for-good
github.com
To view or add a comment, sign in
-
🚀 Excited to share my first project for the Headstarter Accelerator: Customer Churn Prediction!** 🏦✨ This project helps banks predict which customers are at risk of leaving, enabling them to take proactive steps for customer retention. 🔐 Here’s how I approached it: - 🔍 Data Preparation: Processed bank customer data from Kaggle to ensure clean, quality input for model training. - 🤖 Model Training: Trained 5 machine learning models, including XGBoost Random Forest, and **Gradient Boosting**, with a focus on optimizing recall to accurately identify potential churners. - 🌐 Interactive Web App: Developed a web app using Streamlit to visualize model performance. Integrated Llama 3.2 for generating customer-friendly explanations and automated churn-prevention emails. 🖥️ Explore the app here: https://lnkd.in/g3TQm5he - 🛠️ API Deployment: Deployed my API using FastAPI, allowing easy integration of the churn models into any application. 🚀 #MachineLearning #Streamlit #ChurnPrediction #CustomerRetention #FastAPI
Streamlit
hs-proj-1-webapp-2.onrender.com
To view or add a comment, sign in
-
Make your custom GPTs smarter with this hack: Setup: 1 - Connect custom GPT to a Webhook 2 - Connect the Webhook to any application Outcome: Allow ChatGPT to "talk" to your data! Endless possibilities. I showed how to do this in my previous video but didn't go into the details of the Make webhook setup. So I created this video to fill that gap. Link in comments. *** Hit 'follow' to level up your no-code game. ↗ #innovation #software #nocode #future
To view or add a comment, sign in
-
UPDATE 😊 : Finally created a small streamlit app on HuggingFace to using gemma 2b, 7b and it LLMs models along with few EDA. Here is the repo 🤔 : https://lnkd.in/e3Pvd25q Here is the scrapper repo on GitHub 🤗 : https://lnkd.in/eT8FK4xU Use the CSV file in the git repo to upload on the dashboard. 😥 HuggingFace really is a game changer for anyone who wants to update their skills with the LLMs, gonna build more things to answer my why? and when? because let's face it we all know the WHAT's in the world when it comes to problems. 😤
LinkedIn Bot Gemma Streamlit - a Hugging Face Space by negismohit123
huggingface.co
To view or add a comment, sign in
-
Last night I tested a script I made to integrate my local Ollama Model with Twilio for SMS messaging. The goal was to explore replacing Open AI API calls and have more control of the Agent pipeline. It worked like a charm! A simple Flask App with NGrok to expose my local endpoint to a public endpoint that can used to for my overall Metis OS Logic. Open source models are just getting better and I'm excited about building an AI Agent system that will save lives! #metis #AgenticAI
To view or add a comment, sign in
-
Here is a Claude use case. Take a bunch of data and preferences, throw it into Claude and ask for an interactive bot, that can be used to help you make decisions, scoring and reward features and tailor as necessary. Learn more at S2udios.com
To view or add a comment, sign in
-
Building a #search or #RAG #chatbot starts with #connecting your #data using SearchBlox! https://lnkd.in/eVRcGueP
To view or add a comment, sign in
-
Our latest update introduces a feature that enables seamless integration of multiple premade tools and web scrapers into your personalized GPTs on ChatGPT, eliminating the need for coding. Experience the convenience of Apify for enhancing your ChatGPT experience. https://lnkd.in/ePaVvkSY
This platform allows you to add dozens of premade tools and web s
apify.com
To view or add a comment, sign in
216 followers