Abheek Gulati

LARS. An application for running LLMs locally on your device, with your documents, facilitating detailed citations in generated responses.

hf-waitress. Serving LLMs in the HF-Transformers format via a PyFlask API

kosmos-2_5-containerized. Kosmos-2.5 is a cutting-edge Multimodal-LLM (MLLM) specializing in image OCR. However, its stringent software requirements & Python-script based invocation make it difficult to use for application development. Here, it has been containerized and made available via an API, greatly enhancing its ease-of-use.

airtable-to-baserow. A tool to securely migrate data from AirTable to Baserow, rather than generating a publicly accessible link of AirTable data for Baserow to load from, as that risks violating compliance law and security best-practices.

Privion. A privacy-first platform for deploying and operating open‑weight LLMs.

JAVA-TCP-Server-and-Client. Basic Java TCP server and client

Chatbot-Conversation-Logger. Extract logs from a cloud based chatbot via a REST API end-point

llama.cpp. LLM inference in C/C++

TenPin-TakeHome. Python

docling-ocr-tool. Docling OCR Tool to Facilitate Evals - Processes an Entire Folder, Persists Settings & Supports a Large Number of File Formats by Leveraging Unoconv.py to Streamline Around PDFs