Skip to content
View poeticmage's full-sized avatar
🎯
Focusing
🎯
Focusing
45 followers · 49 following

Highlights

  • Pro

Block or report poeticmage

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.

Content in all repositories owned by your account will be closed.
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
poeticmage/README.md

Priyadarshi (poetic mage)

I am Priyadarshi, a Software Engineer with experience building backend systems, APIs, distributed applications, and AI/ML solutions. Experienced with Python, Java, SQL, REST APIs, microservices, databases, and cloud-oriented development, alongside Machine Learning and Deep Learning. Built systems ranging from real-time applications and multi-agent workflows to computer-vision models and end-to-end AI products. Interested in Software Engineering, Backend Systems, AI/ML, and scalable automation.

Resume

Contact

Mail: mookherjipriyadarshi@gmail.com

Linkedin: www.linkedin.com/in/priyadarshi-codes

Projects

Refer Eye Glaucoma: Designed and launched an end to end Computer Vision application for Glaucoma Detection in diabetic patients whose Fundus images are used to classify into Referrable and Non-referrable Glaucoma. • Built a MaxViT algorithm from scratch using Pytorch having Block Attention, Grid Attention and Squeeze-Expansion Blocks. Trained and tested it with over 12000 images from EyePacs dataset. Obtained top-1 accuracy of 82%. Used the trained weights and created a server (Microservice) with FastAPI, deployed in Hugging Face Spaces. • Used pre-trained Resnet-18 and DBSCAN to cluster embeddings of fundus images as an initial gate to reject nuisance input images. Implemented it as a RESTful service (Microservice), deployed in Hugging Face Spaces. • Built and deployed a backend gateway server in Render that connects with the 2 Microservices and the frontend. The frontend takes an input, gives the predicted class and a Gradcam image as output if the input is accepted by the gate. • Implemented Cron-job to keep services awake every 14 minutes. https://funduseyefrontend12.onrender.com/ and the Git Link is this: https://github.com/poeticmage/image (backend) and https://github.com/poeticmage/FundusEyeFrontend (frontend). In the image repository, the full AI algorithms are written in the README file, though they are deployed in Hugging Face Spaces

Dynamic Workflow: Built a Multi-Agent system that produces a modular DAG – Activity on the Vertex Curve, using parallelism and dependency metrics, validates it, executes it and allots external agents for task completion. • Achieved objective within 100 to 300 seconds for a graph of 5 nodes for an “interview task”, from planning to execution with batch routing of external agents, consumed 75k to 100k Gemini tokens. The github link is: https://github.com/poeticmage/Dynamic-Workflow

Pinned Loading

  1. FundusEyeFrontend FundusEyeFrontend Public

    This is the frontend for the Refer Eye Glaucoma app. myreactapp was used to create it. The backend will be found in image repo. It is deployed in render (ci/cd)

    JavaScript 1

  2. image image Public

    This is the backend for Refer Eye Glaucoma app. It handles image for fundus-images, sending them to 2 other apis deployed in Hugging Face spaces.

    JavaScript 1

  3. Notebook-data-pipeline Notebook-data-pipeline Public

    This repository contains jupyter notebooks of data preprocessing of various datasets before being fed to the model. Proper train,test split, class handling and train test loaders are written in cod…

    Jupyter Notebook 1

  4. Deep-Learning-Models-NB Deep-Learning-Models-NB Public

    This repo contains notebooks of various deep learning models

    Jupyter Notebook 1

  5. Dynamic-Workflow Dynamic-Workflow Public

    This is a Modular Dynamic Workflow orchestration with Gemini SDK and Google ADK 2.5 that creates a DAG plan, executes and validates it

    Python 1