Projects

Overview: This repository contains the official code for the book "Hands-On Large Language Models," authored by Jay Alammar and Maarten Grootendorst. The book is available on Amazon, Shroff Publishers, O'Reilly, and other platforms. Key features: - The code for all examples in the book - Visual educational nature with almost 300 custom figures - Available on various platforms including Amazon, Shroff Publishers, O'Reilly, and Google Colab Tech stack: Jupyter Notebook Why it matters: This repository provides practical tools and concepts for using Large Language Models, making it essential for developers looking to understand and implement these models in their projects.

### Overview: OpenDataLoader PDF is a powerful tool for extracting structured data from PDFs, designed to automate PDF accessibility and improve accessibility remediation. It offers benchmark accuracy in PDF extraction and is the first open-source solution for generating Tagged PDFs from untagged PDFs. ### Key Features: - **PDF Extraction**: Extracts Markdown, JSON, and HTML from any PDF. - **Benchmark Accuracy**: Achieves the highest accuracy in benchmarks (0.907 overall), with 0.928 table accuracy across 200 real-world PDFs. - **Hybrid Mode**: Works with poor-quality scans (300 DPI+) and complex tables, formulas, and images. - **Auto-Tagging**: Automatically tags untagged PDFs into accessible Tagged PDFs, making them compliant with PDF/UA standards. ### Tech Stack: Java 11+ ### Why It Matters: - **High Accuracy**: The tool excels in extracting structured data from PDFs, making it valuable for RAG/LLM pipelines and data extraction. - **Accessibility**: It automates PDF accessibility remediation, making it crucial for compliance with accessibility regulations. - **Open-Source**: The project is open-source, ensuring transparency and collaboration in the accessibility community. ### Repository: The repository is hosted on GitHub and is part of the OpenDataLoader project. It has received 29,732 stars, indicating its widespread adoption and importance in the field of PDF accessibility and data extraction.

Overview: Kestra is an open-source event-driven orchestration platform designed for mission-critical applications, providing a declarative, language-agnostic interface for orchestrating data, AI, and infrastructure workflows. Key features: - Event-driven scheduling and orchestration - Infrastructure as Code best practices - AI Copilot for generating workflows - Agent Skills for generating workflows directly from coding agents - Extensible and developer-friendly plugin ecosystem Tech stack: Java Why it matters: Kestra simplifies the orchestration of complex workflows, making it easier to build, manage, and scale reliable applications with minimal manual intervention.

Overview: TencentDB Agent Memory is a comprehensive memory hub designed for AI agents, turning conversations, documents, and code into reusable assets that can be shared across teams and frameworks. Key features: - **Memory Assets**: Chat Memory, Skill, LLM-Wiki, Code-Graph - **Governance**: Assets are governed, shared, and equipped across agents and frameworks - **Team Collaboration**: Supports multiple agents sharing the same memory server - **Integration**: No plugins, hooks, or MCP servers required, just point the Agent's base URL to the Proxy Tech stack: TypeScript Why it matters: TencentDB Agent Memory simplifies the process of integrating AI agents with existing systems, making it easier for teams to collaborate and innovate. Repository: [TencentCloud/TencentDB-Agent-Memory](https://github.com/TencentCloud/TencentDB-Agent-Memory) GitHub stars: 29,507 GitHub description: TencentDB Agent Memory is a team-level memory hub for AI Agents — turning conversations, docs, and code into four reusable memory assets (Chat Memory, Skill, LLM-Wiki, Code-Graph) that are governed, shared, and equipped across agents and frameworks.

Overview: Mr. Ranedeer: Your personalized AI Tutor is a GPT-4 AI prompt designed to provide tailored learning experiences for users with various needs and interests. Key features: - Customizable learning style and communication type - Adjustable depth of knowledge - Personalized reasoning framework Tech stack: The prompt is powered by GPT-4. Why it matters: Mr. Ranedeer offers a highly customized and effective learning experience, making it suitable for users of all backgrounds and learning styles. Repository: JushBJJ/Mr.-Ranedeer-AI-Tutor GitHub stars: 29,552 GitHub description: A GPT-4 AI Tutor Prompt for customizable personalized learning experiences.

Overview: Chroma is an open-source data infrastructure designed for AI, providing a fast, cost-effective, and scalable search solution. Key features: - **Vector Search**: Supports vector search for efficient similarity-based queries. - **Hybrid Search**: Combines vector and full-text search for comprehensive results. - **Full-Text Search**: Handles full-text queries for natural language understanding. Tech stack: Rust Why it matters: Chroma simplifies the development and deployment of AI-powered search solutions, making it accessible and efficient for various applications. Repository: chroma-core/chroma GitHub stars: 29,521 GitHub description: Search infrastructure for AI Docs: https://docs.trychroma.com Roadmap: https://docs.trychroma.com/docs/overview/oss#roadmap Contributing: https://docs.trychroma.com/docs/overview/oss#contributing License: Apache 2.0

### Overview: iFixAi is an independent auditing tool designed to verify if AI agents are performing as intended. It provides real-time insights within 120 seconds, ensuring accountability and transparency in AI agent operations. ### Key Features: - **Real-Time Auditing**: Quickly assess AI agent performance without delays. - **Automated Checks**: Automated checks to identify discrepancies and misalignments. - **User-Friendly Interface**: Intuitive dashboard for easy monitoring and management. ### Tech Stack: - Python ### Why it Matters: iFixAi helps organizations ensure that AI agents are operating ethically and efficiently, preventing costly mistakes and maintaining trust in their AI solutions.

Overview: Page Agent is a JavaScript in-page GUI agent that allows you to control web interfaces with natural language, without the need for browser extensions, Python, or headless browsers. Key features: - Easy integration: No need for browser extensions, Python, or headless browsers. - Text-based DOM manipulation: No screenshots or multi-modal LLMs required. - Bring your own LLMs: Works with most mainstream models, including locally deployed ones. - Optional Chrome extension for multi-page tasks. - MCP Server (Beta) for remote control. Tech stack: TypeScript Why it matters: Page Agent simplifies the integration of AI into web applications, making it accessible and easy to implement without the complexity of traditional solutions. Repository: alibaba/page-agent Primary language: TypeScript GitHub stars: 29,472 GitHub description: JavaScript in-page GUI agent. Control web interfaces with natural language.

Overview: This repository contains a collection of data science Python notebooks focusing on various aspects of deep learning, statistical inference, and machine learning using TensorFlow, scikit-learn, and other tools. It includes tutorials and examples for deep learning, statistical methods, and data analysis. Key features: - TensorFlow tutorials - Scikit-learn tutorials - Statistical inference tutorials - Pandas tutorials - Matplotlib tutorials - NumPy tutorials - Python essentials - AWS tutorials - Command lines Tech stack: Python Why it matters: This repository provides comprehensive resources for learning and applying data science techniques using Python, TensorFlow, and related tools, making it valuable for developers and data scientists looking to enhance their skills in these areas. Repository: donnemartin/data-science-ipython-notebooks GitHub stars: 29,368 GitHub description: Data science Python notebooks: Deep learning (TensorFlow, Theano, Caffe, Keras), scikit-learn, Kaggle, big data (Spark, Hadoop MapReduce, HDFS), matplotlib, pandas, NumPy, SciPy, Python essentials, AWS, and various command lines.

### Overview: This repository is a curated list of awesome Deep Learning tutorials, projects, and communities. It serves as a comprehensive resource for anyone interested in learning about and applying deep learning techniques. ### Key Features: - Tutorials and courses on various aspects of deep learning. - Projects and examples demonstrating practical applications of deep learning. - Communities and forums for discussion and collaboration. ### Tech Stack: Python, TensorFlow, Keras, PyTorch, and other relevant deep learning frameworks. ### Why It Matters: Deep learning is revolutionizing fields such as computer vision, natural language processing, and more. This repository provides a wealth of resources for anyone looking to deepen their understanding and expertise in this field. Whether you're a beginner or an experienced practitioner, this list offers valuable insights and practical tools to enhance your skills.

Overview: This project is a comprehensive book on machine learning systems, covering foundational principles, scaling techniques, agentic AI, and physical AI. It is part of the Harvard CS249r series and is available in multiple languages. Key features: - Foundations of Machine Learning Systems - Scaling Techniques - Agentic AI - Physical AI Tech stack: Python Why it matters: This book provides a deep understanding of the principles and practices of engineering artificially intelligent systems, making it essential reading for developers and researchers in the field of machine learning. Repository: harvard-edge/cs249r_book Primary language: Python GitHub stars: 29,058

Overview: XGBoost is a high-performance, scalable, and flexible library for gradient boosting, designed to be efficient, flexible, and portable. It supports Python, R, Java, Scala, and multiple languages including C and C++. XGBoost is optimized for distributed computing environments, including Kubernetes, Hadoop, Spark, Dask, and Flink. Key features: - Efficient parallel tree boosting - Supports multiple programming languages - Runs on single machine, Hadoop, Spark, Dask, Flink, and DataFlow - Optimized for distributed computing environments - Supports Python, R, Java, Scala, and multiple languages including C and C++ - Provides a scalable, portable, and flexible solution for machine learning problems Tech stack: C++ Why it matters: XGBoost is a powerful tool for machine learning that can handle large datasets efficiently and effectively, making it ideal for applications requiring high performance and scalability.

### Overview: Repomix is a powerful tool that converts your entire repository into an AI-friendly format, making it easy to feed your codebase to various large language models and AI tools. ### Key features: - Automatically packs your entire repository into a single, AI-friendly file. - Supports multiple AI models like Claude, ChatGPT, DeepSeek, Perplexity, Gemini, Gemma, Llama, Grok, and more. - Ideal for developers needing to integrate their codebase with AI tools. ### Tech stack: TypeScript ### Why it matters: Repomix simplifies the integration of your codebase with AI models, making it easier to leverage advanced AI technologies for development and code analysis.

Overview: This project is a comprehensive daily study plan designed to help software engineers transition into becoming a machine learning engineer. Key features: - Detailed daily study schedule - Hands-on approach focusing on practical applications - Abstraction of complex mathematical concepts - Aligns with the "Top-down learning path" method Tech stack: Python, TensorFlow, Jupyter Notebook Why it matters: This plan provides a practical and accessible way for software engineers to learn machine learning, making it easier to apply the skills in real-world projects. Repository: ZuzooVn/machine-learning-for-software-engineers GitHub stars: 288,730 GitHub description: A complete daily plan for studying to become a machine learning engineer.

### Overview: Resume Matcher is a self-hosted AI harness designed for developers, offering a local and open-source solution for creating professional resumes and cover letters. Tailored to each job description, it supports multiple language models including Claude, ChatGPT, DeepSeek, Kimi, GLM, and Gemma. ### Key Features: - **AI-Generated Resume Creation**: Automatically tailors your resume to match job descriptions. - **Multiple Model Support**: Compatible with various language models for flexibility. - **Local and Open-Source**: Runs on your machine, supports local models, and is free and open-source. - **Versatile Layouts**: Customize the resume layout and sections to fit your personal style. - **Job Application Tools**: Includes a cover letter and interview preparation feature. ### Tech Stack: - **Python**: The primary programming language used for the backend. - **FastAPI**: Fast, stateless, and scalable backend framework. - **Next.js**: Frontend framework for building user interfaces. - **SQLite**: Database management system for local storage of resumes. ### Why It Matters: Resume Matcher simplifies the process of creating professional documents tailored to job applications, making it a valuable tool for developers looking to stand out in their field. Its local and open-source nature ensures it can be easily integrated into any existing workflow without requiring an external account or database management. ### Repository: - **Repository**: srbhr/Resume-Matcher - **GitHub Stars**: 28,736 - **GitHub Description**: The #1 AI Harness for Building Resumes, PDFs

### Overview: Invoke is a comprehensive creative engine for Stable Diffusion models, providing professionals and enthusiasts with the ability to generate and create visual media using the latest AI technologies. It offers a user-friendly web interface and serves as the foundation for multiple commercial products. ### Key Features: - **Unified Canvas**: A fully integrated canvas with support for all core generation capabilities, including in/out-painting, brush tools, and more. - **Workflows & Nodes**: A fully featured workflow management solution that allows users to develop and share customizable generation pipelines. - **Board & Gallery Management**: An organized gallery system for storing, accessing, and remixing content in the Invoke workspace. - **Model Support**: Support for various models such as SD 1.5, SD 2.0, SDXL, SD 3.5 Medium, SD 3.5 Large, CogView 4, Flux.1, Flux.2, and more. ### Tech Stack: TypeScript ### Why it Matters: Invoke is a leading creative engine for Stable Diffusion models, empowering professionals, artists, and enthusiasts to generate and create visual media using the latest AI-driven technologies. The solution offers an industry leading web-based UI, and serves as the foundation for multiple commercial products. Repository: [invoke-ai/InvokeAI](https://github.com/invoke-ai/InvokeAI) GitHub Stars: 28,712

### Overview: Semantic Kernel is a model-agnostic SDK that enables developers to quickly integrate cutting-edge LLM technology into their applications, providing enterprise-grade reliability and flexibility. ### Key features: - Connects to any LLM with built-in support for OpenAI, Azure OpenAI, Hugging Face, and more. - Empowers the creation of modular AI agents with access to tools/plugins, memory, and planning capabilities. - Orchestrates complex workflows with collaborating specialist agents. - Extends with native code functions, prompt templates, OpenAPI specs, or Model Context Protocol (MCP). - Seamlessly integrates with vector databases like Azure AI Search, Elasticsearch, and Chroma. - Processes text, vision, and audio inputs. - Runs locally with Ollama, LMStudio, or ONNX. - Supports structured business processes with a workflow approach. - Built for observability, security, and stable APIs. ### Tech stack: .NET, Python, Java ### Why it matters: Semantic Kernel simplifies the integration of advanced language models into applications, making it easier for developers to create intelligent AI agents and multi-agent systems.

### Overview: Spleeter is a Python library and command-line tool designed by Deezer for separating audio sources into distinct components such as vocals, drums, and other instruments. It leverages pretrained models to achieve high-quality separation and is optimized for both real-time and batch processing. ### Key Features: - **2-4 Stems Model**: Separates vocals, accompaniment, and drums into two or four stems. - **4 Stems Model**: Separates vocals, drums, bass, and other instruments into four stems. - **5 Stems Model**: Separates vocals, drums, bass, piano, and other instruments into five stems. - **Command Line Interface (CLI)**: Easy to use from the command line for real-time processing. - **Python Library**: Available as a Python library for integration into development pipelines. - **Docker Image**: Provides a lightweight container for easy deployment. - **High Performance**: Performs well on the MusDB dataset and can handle large volumes of audio files. - **Versatile Usage**: Can be used directly in development pipelines or through a GUI or standalone website. ### Tech Stack: Python ### Why it Matters: Spleeter simplifies the process of source separation, making it accessible and efficient for developers and audio enthusiasts alike. Its pretrained models ensure high-quality separation, and its fast processing capabilities make it suitable for real-time applications and batch processing.

### Overview: Label Studio is a comprehensive data labeling tool designed for various data types, including audio, text, images, videos, and time series. It offers a user-friendly interface for labeling data and supports exporting to multiple model formats, making it an essential tool for improving machine learning models. ### Key Features: - **Multi-Type Data Labeling**: Supports audio, text, images, videos, and time series. - **Standardized Output Format**: Ensures consistent data representation and compatibility with models. - **User-Friendly Interface**: Simple and intuitive design for efficient data labeling. - **Export Options**: Export data to various model formats for model training. ### Tech Stack: - TypeScript ### Why It Matters: Label Studio simplifies the process of preparing and improving data for machine learning models, ensuring more accurate predictions and better model performance. Its standardized output format and user-friendly interface make it a valuable tool for data scientists and machine learning engineers. ### Repository: - Repository: HumanSignal/label-studio - Primary Language: TypeScript - GitHub Stars: 28,480 - GitHub Description: Label Studio is an open-source data labeling tool with a standardized output format for various data types. ### GitHub Description: Label Studio is a multi-type data labeling and annotation tool with a standardized output format. It is designed to help you label data types like audio, text, images, videos, and time series. The tool simplifies the process of preparing raw data or improving existing training data, making it easier to get more accurate ML models. ### Included Templates: Label Studio comes with

Overview: MLflow is an open-source AI engineering platform designed for teams of all sizes to debug, evaluate, monitor, and optimize production-quality AI applications while controlling costs and managing access to models and data. Key features: - Production-grade observability - Production-grade evaluation - Production-grade prompt management - Production-grade prompt optimization - AI Gateway for managing costs and model access Tech stack: Python Why it matters: MLflow enables teams to ship AI to production with confidence, making it a crucial tool for organizations looking to deploy AI applications efficiently and securely.