Hi, I'm Shantanu Gupta, a CSE undergrad at IIIT Naya Raipur. I am a tech enthusiast interested in Machine Learning, Web Development and Data Science.
Experience
Twenty | Open Source Contributor
Contributed to Twenty, an open-source CRM, by enhancing core features, improving code quality through multiple successful pull requests, optimizing the webhook interface with directional flow arrows for better data flow clarity, and resolving a critical bidirectional Unicode rendering bug to ensure cross-platform compatibility.
HuggingFace | Open Source Contributor
Contributed to Hugging Face’s open-source ecosystem by improving core machine learning libraries including Transformers, PEFT, and Accelerate, delivering production-grade fixes and documentation updates through multiple merged pull requests, enhancing numerical stability and mixed-precision training behavior in Transformer rotary embeddings, modernizing fine-tuning examples by replacing deprecated APIs, and resolving experiment-tracking issues in distributed and offline training workflows to improve developer experience and reliability.
Latest Projects
File Sharing Server
A secure file sharing platform built with Golang and React that allows users to upload, manage, and share .txt files through unique public links. Authenticated users can access a personalized dashboard to view and manage their uploads. The backend features JWT-based authentication, file validation, and metrics monitoring via Prometheus and Grafana. Designed for simplicity and security, the system ensures efficient file handling with robust access control.
Blockchain based Energy Trading Platform
A decentralized platform for buying and selling renewable energy using Ethereum blockchain and smart contracts. Users can list, buy, and trade energy securely, with transactions stored on the blockchain. It works on the Sepolia Testnet ETH network.
Spatiotemporal Models for Volumetric MRI to Classify Brain Tumor
This research explores an innovative approach to classifying 3D volumetric MRI scans using spatiotemporal models traditionally applied to video analysis. These models typically process data along three dimensions, handling spatial information across two axes and temporal data along the third using architectures such as ResNet(2+1)D and ResNet3D.
Applicant Tracking System
An ATS system that collects and parses resumes and job descriptions by utilizing NLP and ML algorithms to employ semantic matching providing accurate resume ranking and scoring. It is deployed on a full stack web interface.
Social Media Sentiment Analysis
A sentiment analysis model trained on tweets, YouTube comments and Amazon reviews. It is has a full stack web interface and multi-language support.
Contact Me