Which Open‑Source Projects Won InfoWorld’s 2023 Bossie Awards?

InfoWorld’s 2023 Bossie Awards recognize the most innovative open‑source software across data management, AI, and development tools.

Apache Hudi

Provides fast data formats, tables, and SQL for low‑latency real‑time analytics, integrating with Spark, Flink, Presto, StarRocks, and Amazon Athena.

Apache Iceberg

Offers high‑performance table formats with schema evolution, compression, and version rollback, compatible with Hive, Spark, Flink, ClickHouse, Dremio, and StarRocks.

Apache Superset

A comprehensive data‑visualization platform supporting a wide range of charts, SQL databases, drag‑and‑drop builders, and a SQL IDE for large‑scale self‑service analytics.

Bun

An all‑in‑one JavaScript runtime, bundler, and package manager that serves as a faster alternative to Node.js and NPM, built on Zig.

Claude 2

Anthropic’s large language model that handles up to 100K tokens per prompt, capable of editing, summarizing, extracting data, and answering questions across many languages.

CockroachDB

A distributed SQL database delivering ACID transactions with horizontal scalability, multi‑region and multi‑cloud support, used by companies like Netflix, Hard Rock, JPMorgan, and DoorDash.

CPython

Recent Python 3.11/3.12 releases improve interpreter performance and plan to remove the Global Interpreter Lock, enhancing multi‑threaded parallelism.

DuckDB

An embedded OLAP database offering SQL, ACID, secondary indexes, and support for large‑scale joins and aggregations, with native Parquet handling.

HTMX & Hyperscript

Extend HTML with attributes for AJAX requests and dynamic behavior, reducing JavaScript boilerplate and providing a modern alternative to traditional front‑end frameworks.

Istio

A service mesh for container‑based microservices that provides traffic routing, observability, and security features, graduating from CNCF incubation in 2023.

Kata Containers

A secure container runtime combining container speed with VM isolation, supporting multiple architectures and backed by major cloud providers.

LangChain

A modular framework for building applications powered by language models, with integrations for Python, TypeScript/JavaScript, and Go.

LM‑Eval‑Harness

EleutherAI’s benchmark suite with over 200 tests, used to evaluate and compare large language models.

Llama 2

Meta’s next‑generation LLM, trained on 2 trillion tokens, featuring supervised fine‑tuning and RLHF for safety and usefulness.

Ollama

A command‑line tool for running Llama 2, Code Llama, and other models locally on macOS, Linux, and soon Windows.

Polars

A fast DataFrame library for Python, offering up to ten‑times speed improvements over Pandas with lower memory usage.

PostgreSQL

Version 16, released in September 2023, adds performance enhancements, parallel query support, finer‑grained security, and integration with vector extensions for generative AI.

QLoRA

A quantization technique enabling fine‑tuning of 30B‑plus parameter models on a single desktop with minimal accuracy loss.

RAPIDS

GPU‑accelerated libraries for data science, including cuDF, cuGraph, and cuML, with integrations for Spark, SQL, and Dask.

Spark NLP

A natural‑language‑processing library for Apache Spark supporting over 20 000 models for tasks like translation, NER, and sentiment analysis.

StarRocks

A high‑performance query engine combining native C++ code, cost‑based optimization, and SIMD vector processing for real‑time analytics on data lakes.

TensorFlow.js

Brings TensorFlow’s machine‑learning capabilities to JavaScript, enabling model execution in browsers and Node.js with WebGL acceleration.

vLLM

An inference engine delivering up to 20× higher throughput than Hugging Face Transformers, supporting OpenAI‑compatible APIs and FlashAttention 2.

Weaviate

A vector database that combines keyword and vector search, offering modules for retrieval‑augmented generation and multimodal data handling.

Zig

A systems programming language aiming to replace C with safer syntax and performance‑oriented memory control.