/curated-projects

Curated list of Industrial Projects for developers and students to replicate and learn.

Cloud and ML Based Project Ideas

1. SME Chatbot

Domain: Education

Tags: Serverless, GenAI

Technology: Azure OAI, OpenAI

Complexity: Beginner

Description: Create an AI SME (Subject Matter Expert) which will help students of different streams to address their day to day queries, help them in assignments in progressive way. But, it should be responsible AI, filter only relatable and specific content, and restrict the rest of the content.

Related Links:

2. Cloud Video Streaming Platform

Domain: Online Streaming

Tags: Serverless, Video Streaming

Technology: Azure AMP Embeddings, Vimeo

Complexity: Beginner

Description: Create a realtime Video streaming platform, application can be hosted on Cloud. Understand the streaming technology behind Netflix, Hotstar, etc., storing videos on Cloud Object Storage, and host it on Website UI.

Related Links:

3. Image Analysis Application

Domain: Image Detection and Identification

Tags: Machine Learning (Computer Vision)

Technology: GCP AutoML Vision, Azure ML Studio

Complexity: Intermediate

Description: Create a project to derive insights from your images in the cloud or at the edge with AutoML Vision or use pre-trained Vision API models to detect emotion, understand text, and more. Just upload the image, and let ML do the rest

Related Links:

4. Secure file storage with Hybrid Cryptography

Domain: Security

Tags: Cloud Security

Technology: AWS S3, Azure File Storage

Complexity: Intermediate

Description: Create a project to store file in completely secured manner, as the file is being encrypted not by just using one but three encryption algorithm which are AES, DES and RC6. Application ensures security of the user’s data stored on cloud by providing a tool that helps to encrypt files. The user must receives the key via email.

Related Links:

5. Realtime Social Media Sentiment Analysis

Domain: Sentiment Analysis

Tags: Machine Learning (NLP)

Technology: GCP AutoML , Azure ML Studio

Complexity: Intermediate

Description: Create a project for the Sentiment Analysing to predict the emotional state of an individual based on the data input from the person's social media posts and interactions. To perform this analysis, integrate platform like twitter, instagram, etc., to the website or the mobile application and take the user authorized account name and process the data accordingly in your backend server and finally display the result of analysis performed i.e. the emotional state of the individual whose account name was provided.

Related Links:

6. FaceNet Distributed Training with Kubeflow

Domain: Image Detection and Identification

Tags: MLOps

Technology: Kubeflow, OpenCV, Kubernetes

Complexity: Hard

Description: Create a project to demonstrates how to use kubeflow end-to-end to use Tensorflow MuiltiWorkerMirroredStrategy on the local Kubernetes cluster for distributed training, and implement image similarity calculation through FaceNet algorithm and OpenCV face recognition technology.

Related Links:

7. Cloud-Based E-Commerce Platform

Domain: E-Commerce

Tags: Full Stack

Technology: Cloud Database (NoSQL), Containers

Complexity: Hard

Description: Create a project which demonstrates e-commerce utility. You need to use Databases hosted in cloud environments for storing user information, product details, admin information.

Related Links:

8. Phishing platforms identification Portal

Domain: Security

Tags: Machine Learning, Data Mining

Technology: Deep Learning, Azure threat protection ML

Complexity: Hard

Description: Create a project will identify whether the website link is phishing or scam. Break entire application in THREE parts:

  • Website: Create a webapp for interacting with ML models using API.
  • ML model: Use Deep Learning algorithms for building advanced AI based phising link detection system.
  • Cloud: Use cloud to host/deploy your ML model and use cloud based fast, realtime, end-to-end encrypted APIs to interact with the model.

Related Links: