Founding Engineer @ Fable | Ex. Eaton Aerospace | Speaker | VIT '23

projects

hashnode

youtube

Kedar Basutkar  • Peerlist

Project: VideoAI - a tool that lets users search video content using natural language queries, similar to querying a database.Purpose: Eliminates the need to manually scrub through video footage by enabling users to retrieve specific moments from videos through simple queries like “show me the crocodile.”Key Features:<ul><li>Allows users to query video content using natural language.</li><li>Provides precise timestamps of relevant video frames based on the query.</li></ul> How VideoAI Works: 1. Video Processing Flow: - Frame Extraction: Extracts frames from the uploaded video at regular intervals (e.g., every 1 second). - Image Captioning: Uses an image captioning model to generate descriptions for each frame (e.g., "a crocodile is swimming"). - Storage in Vector Database: Captions and timestamps are stored in a vector database for efficient search and retrieval. 2. Video Querying Flow: - Natural Language Query: Converts user queries into vectors using the same model that generated the captions. - Vector Search: Compares the query vector with stored caption vectors to find matches. - Resulting Timestamps: Returns exact timestamps of relevant frames where the query matches.Real-World Applications: 1. Surveillance Footage Analysis: Search for specific events in security footage like "person entering the building." 2. Media Curation for Broadcasters: Helps broadcasters curate specific moments (e.g., “best goals”) from long video footage. 3. Video-Driven Search Engines: Potential to revolutionize video search by allowing users to search based on the actual video content, not just titles or metadata.- Technical Stack: - Runs locally, using AI models for captioning and embeddings. - Stores video data in a vector database for fast search and retrieval.

VideoAI: Query Videos like a Database!

ConvoDigest - an app designed to summarize and query unstructured group chats. <ul><li>Purpose: Helps users in large communities stay on top of long, unstructured group chats by generating summaries and answering queries based on chat content.</li></ul>Key Technologies:<ul><li>AI/LLM Integration: Utilizes large language models (LLMs), retrieval-augmented generation (RAG), and embeddings.</li><li>Vector Database: Employs Vectra for storing chat embeddings.</li><li>LLaMA 3.1 (7B parameters) via Ollama for AI-driven analysis.</li></ul> Key Problem:<ul><li>Managing and retrieving important information from overwhelming group chat volumes (e.g., WhatsApp chats).</li></ul>Solution:<ul><li>Summarizes chat content and allows users to query chats for key points and specific questions.</li><li>Local-first approach: Entire processing is done locally on the user’s device to maintain privacy and security.</li></ul>How It Works:<ul><li>Users upload their WhatsApp chat data.</li><li>App processes the chats using Node.js, Express, and React, generating searchable summaries.</li></ul> Privacy-Focused: No offloading to external servers; all data is processed locally to maintain full user privacy.

ConvoDigest: Local-first AI solution to manage community

Software Engineer | Solving problems through code

Software Engineer - Intern @Fable |Full stack developer  | COEP 25

<h2 class="rich-heading">LearnTube - Unlock the Power of YouTube Learning with Our Course-Boosting App! 👩‍🎓</h2><h2 class="rich-heading">About 🚀</h2>LearnTube is a free online education app that allows learners to access high-quality course playlists on YouTube and track their progress with an analysis dashboard. The app provides learners with data on their learning progress, including time spent on each video, quiz scores, and overall progress. This analysis helps learners stay motivated and accountable for their learning goals.<h3 class="rich-heading">Features ✨</h3><ul><li>Seamless onboarding experience for users</li><li>Courses divided into categories for easy browsing and enrollment</li><li>Analysis graphs with a dashboard for tracking learning progress</li><li>Personalized course recommendations based on learners' interests and learning styles (coming soon)</li><li>Social features for connecting with other learners and sharing progress</li></ul><h3 class="rich-heading">Technologies Used 💻</h3><ul><li>React for the front-end</li><li>Node.js for the back-end</li><li>YouTube's API for course playlist integration and video/channel data retrieval</li></ul>

Learntube🚀: Make the Most of YouTube Courses with LearnTube

Get Interview Ready is the one-stop solution for helping you with all aspects of interview preparation, from practicing technical and behavioral questions to helping you manage your referral network, and so much more!

Get Interview Ready!

The design of leaf springs consists of multiple non-linear equations that have to be solved simultaneously to get an adequate design. 
Used genetic algorithms, which are promising techniques to find the optimal solutions to these equations for weight or stress minimization.

Constrained Design Optimization of Leaf Spring using Genetic

Roadmap Runner is an app that simplifies the process of learning software development. Our app offers a range of features designed to make the learning process more manageable, efficient, and effective.

Some of these features include:
👨‍💻 Expert-created roadmaps for various fields of software development
📈 Progress tracking: users can track their progress through the expert-created roadmaps
📊 Consistency analysis: users can view consistency analysis of their learning process
🏆 Leaderboard feature: users can stay motivated through the leaderboard feature, which allows them to see how they are doing compared to other users.
🙌 Social sharing: users can share their progress on Twitter, making it easy to stay accountable and stay connected with other learners.

With all these features, Roadmap Runner is a comprehensive solution for anyone looking to learn software development in an efficient and effective way.

RoadmapRunner - Navigate the path to success!

devCase helps developers build their portfolio website to showcase their skills to recruiters in minutes. It leverages users' GitHub, Twitter, and Hashnode (blogging website) to automatically build a portfolio.
Built using Next.js, Twitter API, GitHub API, and Hashnode API.
devCase is one of the winners at Crossknot Hacks Hackathon.