Skip to content
View pablopupo's full-sized avatar

Block or report pablopupo

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
pablopupo/README.md

Hi, I'm Pablo

I'm a CS student at the University of Florida and I work in applied AI. Right now I'm doing an AI engineering internship at Handtevy, building document intelligence pipelines for emergency medicine software used by 200,000+ clinicians nationwide, and contributing to open source projects like docling, vLLM, SGLang, and the MCP SDKs.

I'm also a classical pianist. I run Accordo, a booking and payments marketplace for musicians, and I'm building Gradus ad Parnassum, which reads sheet music and answers a musician's questions about it.

Open source

Merged:

  • docling #3702 - optional-dependency imports that broke the core converter in slim installs
  • docling #3721 - code language detection for parsed code blocks

In review:

Projects

  • kit-ai - offline-first emergency first-aid PWA I built with my hackathon team. Llama 3.2 runs on-device through WebLLM and WebGPU. My pieces were the IndexedDB retrieval layer, the online/offline TTS fallback, and a fine-tuned Llama 3 medical model that I'm now working to ship in the app.
  • Accordo - live marketplace for musicians: booking, contracts, payments.

Pinned Loading

  1. gradus-ad-parnassum gradus-ad-parnassum Public

    Gradus ad Parnassum. RAG over musical notation: parses scores, annotates them like a musician would, answers theory questions with measure references.

  2. kit-ai kit-ai Public

    Forked from brandonjacobson/KIT-AI

    Offline-first emergency first-aid PWA. Llama 3.2 runs fully on-device via WebLLM and WebGPU, grounded in a vetted knowledge base, so it keeps working with zero connectivity.

    JavaScript

  3. notes notes Public

    Rough technical notes, updated as I work