Jayakumar Pujar Claim this profile

🤗 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.

Conducted advanced geospatial data analysis using NAIP/NASA imagery for flood detection, building footprint extraction, and vegetation mapping with ArcGIS for deep learning object detection models and multi-layer vector-raster processing.

Transformers-compatible library for applying various compression algorithms to LLMs for optimized deployment with vLLM

AI Agent Governance Toolkit — Policy enforcement, zero-trust identity, execution sandboxing, and reliability engineering for autonomous AI agents. Covers 10/10 OWASP Agentic Top 10.

Weakly supervised deep learning for 1m wetland mapping using multi-temporal NAIP imagery and LiDAR in the Prairie Pothole Region. Built on the geoai open-source package.

Vision Mamba for Image Classification

EHR by using blockchain technology(Hyperledger-Fabric)

It gives the automatic attendance by using Tensorflow.js library to run Face detection and recognition in a video conference application.

test heroku app

Training new TR detection model using Tesseract OCR engine 5.2 with new fonts.

No description.

Face Recognition web application for students attendance.

No description.

Engineered a domain-specific Financial LLM by distilling reasoning traces from DeepSeek-R1 (671B) into Qwen-7B using the Open-R1 methodology; implemented an SFT pipeline on the distilled corpus to replicate complex analytical behavior, achieving near-671B performance with 99% lower inference costs.

Introduction to Agent Development Kit for beginner.

Designed an AI agent using Google’s Agent Development Kit (ADK) with Gemini 2.0 Flash, integrating an Airbnb Model Context Protocol (MCP) server to standardize LLM interactions with external data services and enable autonomous.

GeoAI: Artificial Intelligence for Geospatial Data

Dense Associative Memory for Image Classification with CNN (Resnet50)

OpenCV Text Detection (EAST text detector) OCR

E-voting using Blockchain Technology

Binary or Multi Classifier to classify images by using Deep learning Architecture.

Descentralized voting application using ethereum.

No description.