Thanks to visit codestin.com
Credit goes to github.com

Skip to content
@eulogik

Eulogik

Product development specialists, for startups.

Eulogik 🧠

AI Infrastructure & Quantitative Systems Engineering

Hugging Face Org Website


We build machine intelligence that runs locally, trades systematically, and scales efficiently.
Eulogik is a deep-tech engineering lab. We specialize in tiny, fast, edge-deployable foundation models, system-level optimizations, and quantitative execution infrastructure.


🛠️ Open-Source Repositories

🔬 Edge Models & TinyML

  • TinyDoc-VLM: 256M-param document Vision-Language Model. SigLIP + SmolLM2, LoRA fine-tuning, ONNX export, optimized to run locally on CPU.
  • NanoForecast: Zero-shot time series foundation model family (200K - 6.5M parameters) designed for edge-device streaming inference.
  • pico-type: A 1.5M-parameter, tokenizer-free byte-level classifier (~200KB ONNX, <6ms) for 62 programming languages, 30 text languages, and 90 MIME types.

🐡 Routing & Agent Infrastructure

  • fugusashi: Federated Learning of LLM Routing with human-interpretable decisions. An open-source alternative to Sakana AI Fugu.
  • nexus: Desktop application for local autonomous coding agents that build, run, and debug code from first principles.
  • AutoPR: AI-driven workflow tool for generating automated pull request descriptions and code reviews in seconds.
  • ShellWise: Context-aware terminal copilot that runs locally to assist with shell operations and scripts.

⚖️ Domain-Specific Intelligence

  • LexRAG: Legal intelligence terminal featuring hybrid RAG systems customized for Indian and UAE law, taxation, and compliance.

🏎️ Performance Philosophy

Characteristic Our Philosophy Cloud LLM Status Quo
Deployability local / on-device / CPU / Raspberry Pi Massive server farms / A100/H100 clusters
Inference Latency Sub-10ms (compiled to native C/WASM) 500ms - 2s (network & queue overhead)
Data Privacy Zero data leaves your perimeter External APIs, regulatory compliance risk
Operating Cost Fixed compilation & local runtime cost Variable token pricing with high recurring bills

🤝 Build With Us

We collaborate with forward-thinking engineering teams, startups, and quantitative desks:

  • Consulting & Integrations: Deploying custom edge LLMs, private RAG engines, and automated backtesting engines on your target hardware.
  • Venture Engineering: Spinning out specialized vertical products (like LexRAG) built on top of the Eulogik core stack.

📬 Reach out: [email protected] | eulogik.com/contact

© 2026 Eulogik. Engineered for efficiency.

Popular repositories Loading

  1. fugusashi fugusashi Public

    Fugusashi — Federated Learning of LLM Routing with Human-Interpretable Decisions. Open-source alternative to Sakana AI Fugu.

    Python 5

  2. AutoPR AutoPR Public

    AI-Powered PR Descriptions & Code Reviews in Seconds

    TypeScript 2

  3. pico-type pico-type Public

    A tiny byte-level multi-head content classifier (~1.5M params, ~200KB ONNX, <6ms). Classifies code, text, markup, config, images, binary, secrets, 62 code languages, 30 text languages, 90 MIME type…

    Makefile 2

  4. ShellWise ShellWise Public

    your local AI terminal co-pilot

    Python 1

  5. buffbrain buffbrain Public

    The clipboard manager with a brain.

    Rust 1

  6. NanoForecast NanoForecast Public

    World's most deployable time series foundation model — 200K-6.5M params, zero-shot forecasting, streaming RNN inference, ONNX edge deployment, runs on Raspberry Pi

    Python 1

Repositories

Showing 10 of 11 repositories

Top languages

Loading…

Most used topics

Loading…