Search AI & ML Showcase

Open-source vector similarity extension for PostgreSQL

pgvector adds vector similarity search to PostgreSQL with support for exact and approximate nearest neighbor search. It enables teams to store embeddings alongside relational data without a separate vector database.

Production-ready LLM framework for building RAG and search pipelines

Haystack by deepset is an open-source Python framework for building production NLP and LLM applications. It provides a pipeline abstraction for RAG, document search, question answering, and conversational AI with 60+ integrations.

Unified API for 100+ LLMs — OpenAI-compatible proxy

LiteLLM provides a unified OpenAI-compatible API interface across 100+ LLM providers including OpenAI, Anthropic, Gemini, Mistral, and self-hosted models. It includes a proxy server, cost tracking, rate limiting, and load balancing.

Add validation, correction, and guardrails to LLM outputs

Guardrails AI is an open-source Python framework for specifying and enforcing structural, type, and quality constraints on LLM outputs. It provides a declarative way to validate, retry, and correct AI responses in production.

The industry standard for programmatic workflow orchestration

Apache Airflow is an open-source platform for authoring, scheduling, and monitoring data and ML pipelines as directed acyclic graphs (DAGs). It is the most widely deployed workflow orchestration tool for data engineering teams.

Data framework for building RAG and agentic AI applications

LlamaIndex (formerly GPT Index) is an open-source data framework for building LLM-powered applications. It provides data connectors, indexing, retrieval, and query engines purpose-built for RAG, agents, and knowledge assistants.

Generate massive amounts of realistic fake data for testing and AI

Faker is the most widely used open-source library for generating realistic synthetic data. Available for Python, JavaScript, Go, and more, it generates names, addresses, dates, financial data, and 200+ data types for testing AI pipelines and training data.

Programming framework for algorithmically optimizing LM prompts

DSPy is a Stanford research framework that replaces manual prompt engineering with programmatic optimization. It treats prompts and LM calls as declarative modules and automatically optimizes them using optimizers like BootstrapFewShot and MIPRO.

High-throughput LLM inference engine with PagedAttention

vLLM is an open-source, high-performance LLM inference and serving engine. It uses PagedAttention for efficient KV-cache management, achieving 24× higher throughput than Hugging Face Transformers for production serving.

Framework for orchestrating role-playing, autonomous AI agents

CrewAI is an open-source framework for orchestrating multi-agent systems where agents take on roles, share goals, and collaborate to complete complex tasks. It enables building autonomous teams of AI agents with minimal boilerplate.

DataTruth is an enterprise-grade analytics platform

DataTruth is an enterprise-grade analytics platform that enables anyone to query databases using plain English. No SQL knowledge required—just ask questions and get instant, accurate answers with beautiful visualizations.

A comprehensive, modular Java Machine Learning Framework inspired by scikit-learn, developed by the SuperML community.

SuperML Java 3.1.2 is a sophisticated 21-module machine learning library for Java that delivers enterprise-grade performance with 400K+ predictions/second and 21/21 modules compiling successfully. Now published on Maven Central.