aryandas2911/DailyForge
DailyForge is an open-source fullstack MERN productivity app that lets you design, manage, and visualize your weekly routines - with drag-and-drop scheduling, a smart task library, and overlap protection built right in.
I love to code and am into ML and AI
Repositories
aryandas2911/HemoLens
HemoLens - AI-powered preliminary anemia screening prototype (Grand Hack IPEC 2026): lower-eyelid (+ optional nail) photo → ML Hb estimate → WHO-2024 risk tier → Gemini explainable report.
aryandas2911/DSA-Questions
A personal collection of LeetCode problem solutions solved to strengthen Data Structures & Algorithms fundamentals. Every solution is written in C++, tested locally, and organized by topic and difficulty for easy revision.
aryandas2911/CASCADE
aryandas2911/Agentic-Research-Assistant
A multi-agent AI system that autonomously plans, researches, critiques, and synthesizes comprehensive answers to research questions. Built with Groq, SerpAPI, and Streamlit.
aryandas2911/ParakhAI
AI-powered legal metrology compliance engine, automates packaged commodity inspection via OCR, rule validation, and digital report generation for enforcement officials. (SIH 2026 Hackathon Project Incomplete)
aryandas2911/Akashvaani-AI
Akashvaani AI is an agentic AI-powered GovTech platform that helps citizens discover and apply for government schemes by analyzing documents, checking eligibility, and auto-generating applications. Built at HackMSIT 1.0 (10–11 April 2026) to simplify access to public welfare schemes using AI.
aryandas2911/AI-Tool-Agent
A tool-using AI agent built with LangChain and Groq, exposed through a Streamlit chat interface. The agent decides when to call tools on its own - ask about the weather, give it math to crunch, or request up-to-date information from the web.
aryandas2911/GraphIQ
Turn documents into an interactive knowledge graph. Upload a PDF or DOCX, and GraphIQ extracts entities and relationships from it using an LLM, then lets you explore the result as a force-directed graph
aryandas2911/Hybrid-Search-RAG-with-Pinecone-DB
A comprehensive miniproject demonstrating Retrieval-Augmented Generation (RAG) using hybrid search capabilities with Pinecone, LangChain, and modern NLP embeddings.
aryandas2911/BERT-Question-Answering-Model
A hands-on demonstration of extractive question answering using BERT-large, fine-tuned on the SQuAD v1.1 dataset, built with the Hugging Face transformers library. Given a natural-language question and a passage of text, the model identifies the contiguous span of words in the passage that answers the question.
aryandas2911/Llama-3.2-Fine-Tuning
Fine-tuning Llama 3.2 3B Instruct with QLoRA + Unsloth on math reasoning data (R1-Distill-SFT) to produce a reflective step-by-step reasoning assistant.
aryandas2911/MiniGPT-Text-Generator
A character-level language model implemented in PyTorch, trained on Shakespeare's text (Tiny Shakespeare dataset). Inspired by Andrej Karpathy's "Let's build GPT" series
aryandas2911/AI-Text-Summarizer
A simple web app that summarizes text and documents using pretrained Hugging Face transformer models, wrapped in a Streamlit UI.
aryandas2911/AI-ML-dept-25-26
aryandas2911/aryandas2911
A sleek, dynamic GitHub profile that showcases my journey as a Full-Stack Developer and AI/ML Engineer - built to highlight projects, skills, and relentless growth. Blending code, creativity, and consistency - this is where I document the grind and build in public.
aryandas2911/BERT-Fine-Tuning-for-Text-Classification
Fine-tuning a BERT (bert-base-uncased) model for binary text classification to detect fake news with over 99.5% validation accuracy.
aryandas2911/English-to-French-Translator-Seq2Seq
A Seq2Seq LSTM-based Neural Machine Translation system that converts English sentences into French without attention, built as a clear baseline to understand core translation mechanics.
aryandas2911/IMDB-Dataset-NLP-Model
Built an end-to-end sentiment analysis pipeline on the IMDB dataset, benchmarking classical NLP techniques (OHE, BoW, TF-IDF, Word2Vec) against an LSTM-based deep learning model. Achieved best performance (89.2% accuracy) with TF-IDF + bigrams, highlighting the trade-offs
aryandas2911/SkinSync
SkinSync is an AI-powered skincare analysis web app that scans product ingredient labels and checks compatibility with a user’s skin profile using local vision and language models. Built using the RunAnywhere SDK for HackXtreme Hackathon at GeeksforGeeks, Noida, it runs AI directly in the browser for private, real‑time ingredient intelligence.
aryandas2911/Zorvyn-Assignment-Internship
Interactive finance dashboard built with React, Tailwind CSS, and Recharts for tracking transactions, visualizing spending patterns, and demonstrating role-based UI behavior.
aryandas2911/Stock-Price-Prediction-Model-Stacked-LSTM
Stacked LSTM model for Apple stock price prediction using historical time-series data. RNN LSTM model trained for practice purposes
aryandas2911/CIFAR-100-Classifier-Model-CNN
A CNN project for classifying images from the CIFAR-100 dataset into 100 classes, with data preprocessing, evaluation and prediction.
aryandas2911/Heart-Disease-Prediction-Model-ANN
Artificial Neural Network model for predicting heart disease risk using clinical patient data, covering data analysis, preprocessing, and model evaluation. (Tensorflow/Keras used)
aryandas2911/Indi-Mon
Indi-Mon (CodeZen 2.0 Hackathon @ GTB4CEC): A geospatial AI platform that captures, validates, and stores structured heritage site data in real time, creating a scalable, reliable digital registry of India’s built heritage.
aryandas2911/Next-Arc
A content-based anime recommendation system using TF-IDF vectorization and cosine similarity over anime genres, tags, and themes. Deployed with Streamlit to help users discover their next anime based on what they already love.
aryandas2911/Enginow-Internship-Project-3
Churn Prediction with XGBoost: End-to-end ML pipeline to predict customer churn. Includes feature engineering, class imbalance handling, model training, hyperparameter tuning, and evaluation for actionable business insights.
aryandas2911/Enginow-Internship-Project-2
Credit Risk Prediction System: An end-to-end machine learning project that predicts high-risk vs low-risk loan applicants using XGBoost, SMOTE, and feature engineering, deployed as a public Streamlit web app
aryandas2911/FloPilot
A React-based frontend prototype built for the Snow Frost Hackathon (Jamia Hamdard University) that lets users explain workflows in plain English and visualizes them as real-time flowcharts, demonstrating how AI-guided process clarity transforms operational chaos into professional-grade workflows for early-stage teams.
aryandas2911/Notes-App
📝 Notes App built with the MERN stack (MongoDB, Express, React, Node.js). Features secure user authentication (JWT), create/update/delete notes, and real-time toast notifications for all actions.