Open Source AI Hackathon 2024

GenAI makers and creators contest and showcase

Tickets

Loading…

Amir Nagri

@amir

Bodhilib - An OSS multi-language, performant, plugin architecture based, composable LLM library

Submitted Jan 18, 2024

Bodhilib

Bodhilib is an Open-Source, performant, plugin architecture based, composable LLM library.

Open Source

Source: https://github.com/bodhisearch/bodhilib

Multi-language

  • Bodhilib components are written in Rust
  • Currently supports Python using PyO3
  • Plan is to have language bindings for TypeScript, Java, Swift, WebAssembly
  • Same core, available on multi-platforms, with compatibility
  • Allows applications to be written for Web, Desktop, Mobile, WebAssembly using language of your choice

Performant

With core components written in Rust, bodhilib provides a performance edge over interpreted languages like Python.

Plugin Architecture

  • Bodhilib only defines the core interfaces and models
  • Implementation by vendors
  • Integrated using a Dependency Injection component that supports auto-discovery

Composable

  • The interfaces are designed with composability in mind
  • Easy to create pipelines without need for separate/additional features (like LCEL)

If you are interested in contributing to Bodhilib, sign-up, and I will be very happy to onboard and get you started.

Thanks.

Comments

Login to leave a comment

  • A

    Akshobhya

    @akshobhya_j Editor & Promoter

    @amir, thank you for your proposal submission to The Fifth Elephant Open Source AI Hackathon. The description effectively communicates Bodhilib's vision as an open-source, performant, and composable LLM library with a focus on multi-language support and plugin architecture. Bodhilib's open-source nature is clearly emphasized, supported by providing an easily accessible GitHub repository link. This submission needs to be updated based on the following considerations.

    Concept Evaluation

    The idea of an open-source, performant, multi-language, composable LLM library addresses a relevant need in the development community. The scope and potential impact are substantial, especially with the support for multiple platforms and languages.

    Technical Feedback

    The choice of Rust for core components and Python support with PyO3 demonstrates a strong technical foundation. It's essential to ensure seamless integration and performance optimization for the planned language bindings, especially for mobile and WebAssembly platforms. Consider addressing potential challenges when extending to TypeScript, Java, and Swift.

    Roadmap Assessment

    A clearly defined roadmap detailing the integration plans for additional language bindings and platform support, along with milestone deliverables, would provide transparency and direction for potential contributors.

    Plugin Architecture

    The emphasis on defining core interfaces and models while allowing diverse implementations through a plugin architecture is a sound approach. However, clear documentation and examples will be vital to encourage and guide potential vendors in their implementations.

    Composable Design

    The focus on composability within interfaces is a strong point. Detailing specific use cases and scenarios where this design facilitates pipeline creation without additional features will help potential users understand the value of this approach.

    Community Interaction

    Encouraging contributors to sign up and join the project is a positive step. Providing detailed contribution guidelines and fostering a welcoming environment for new members will be pivotal in building an active and vibrant community around Bodhilib.

    Submission Enhancement

    Consider enriching the proposal with visual aids or diagrams to illustrate the architecture and design concepts. Furthermore, detailing how individuals can contribute, including a breakdown of specific areas where assistance is needed, will attract potential contributors.

    Feedback

    Overall, the Bodhilib proposal holds significant promise. We are enthusiastic about the potential impact of this project and look forward to seeing its development. The inclusivity and technical robustness of the library are commendable, and we encourage you to continue fostering a collaborative and open environment for potential contributors.

    Posted 1 year ago
Hybrid access (members only)

Hosted by

The Fifth Elephant hackathons

Supported by

Host

Jump starting better data engineering and AI futures

Venue host

Welcome to the events page for events hosted at The Terrace @ Hasura. more

Partner

Providing all founders, at any stage, with free resources to build a successful startup.