Kimon N.

From-scratch reimplementation of DeepSeek's Native Sparse Attention (arXiv:2502.11089) in Triton + CUDA Hopper WGMMA. 7.07x faster than FlashAttention-3 at 64k context. Five-model training fleet, perplexity sweep, LongBench v2, MoBA comparison.

Implemented and benchmarked LLM inference compression: int4/int8 quantization, GPTQ-like calibration, int8 KV cache, pruning, distillation, speculative decoding, torch.compile, and ONNX. Every number from a logged, self-audited run.

Tensors and Dynamic neural networks in Python with strong GPU acceleration

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

Code Editor for the AI Agents Era - Run an army of Claude Code, Codex, etc. on your machine

🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.

Microservices TSDB in Go (gateway / ingestor / storage / querier / compactor) with Gorilla compression, PromQL, consistent-hash sharding, and a real-time React dashboard.

No description.

The Postgres development platform. Supabase gives you a dedicated Postgres database to build your web, mobile, and AI applications.

The library for web and native user interfaces.

A framework for building native applications using React

The agent engineering platform.

Matching engine, UDP multicast feed, TCP order gateway, and a React dashboard with order book, depth, heatmap, OHLC chart, and live simulator.

Numerics for white noise and the stochastic heat equation on the torus. Python API, C++/pybind11 hot kernels, analytic Monte Carlo tests.

MLX: An array framework for Apple silicon

End-to-end SCADA fault detection and RCA on Penmanshiel wind-farm telemetry: TimescaleDB hypertables, dbt transforms, classical EWMA baseline plus LSTM autoencoder, SHAP-based attribution, walk-forward CV, Streamlit demo.

TabICLv2: A state-of-the-art tabular foundation model

A self-specializing AI agent that differentiates for code-quality review, security audit, and PR-description quality. Evaluated with 3-seed bootstrap CIs against handcrafted benchmarks and per-domain cross-check oracles that drive rollback.

Parallel HTTP file downloader in Kotlin. N ranged GETs in flight, atomic-assembly + resume, Retry-After honoring, mid-download change surfaced as a typed ValidatorMismatch, single-GET fallback, sha256 & rate-limit. RFC 9110 compliance matrix; Lincheck-verified. Runtime: kotlinx-coroutines-core only. CI: ubuntu/macos/windows × JDK 17/21.

A high-throughput and memory-efficient inference and serving engine for LLMs

Koog is a JVM (Java and Kotlin) framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems

Run Kubernetes locally

The official repository of Mozilla's Firefox web browser.

No description.

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

Straightforward, type-safe argument parsing for Swift

The Swift Programming Language

Production-Grade Container Scheduling and Management

Context retrieval for AI agents across apps and databases