Skip to content

Introduction

Sashi Kumar edited this page May 13, 2019 · 1 revision

What?

Forli is a simple, modern forum that intelligently categorizes questions, suggests the right answers, and routes discussions to the person best capable of answering them. When users ask a question, it's category is instantly determined, and the most relevant answers from related questions are surfaced. Participants in the forum who provide consistent best answers become experts on that category. Questions from those categories are surfaced for them ‘automagically’. Forli also naturally attract angry questions so our sentimental analysis would route such questions to the most empathetic expert capable of answering them.

Why?

The latest iteration of online communication may be social media, but the most popular avenue is still the old-fashioned internet forum. Although there has been a lot of advancements in the field of web technology, forum software systems are still relatively outdated. One of the key issues of today’s forum software is the non-availability of intelligent automation to cut down redundant tasks.

How?

The best way to solve the problems that today’s forums have, is to re-implement the way the forums are built with cutting edge technology. With the rapid growth and inculcation of Machine Learning and NLP algorithms in almost all domains of technology, we believe the next generation forums must be built with ML at its core.

Features:

  • Ability to ask, answer questions. Tag users, comment on questions & answers.
  • Frank, a voice-enabled smart assistant that performs actions on voice command.
  • Experts and moderators for categories.
  • Powerful search and suggestions of related questions.
  • Suggestion of questions to active users to answer them.
  • Automatic categorization of questions.
  • Sentiment analysis techniques to determine the users who need immediate help with answers.
  • Spam Filtering techniques to remove unwanted content getting into the system.
  • Advanced Automation rules to reduce mundane redundant work.
  • Social Media integrations that convert any query related to the configured keywords into a forum question.

Tech Stack:

Backend:

  • Rails (5.1.1)
  • PostgreSQL (9.5)
  • Redis
  • RabbitMQ
  • Elasticsearch
  • Tornado (Python Web App)
  • Scikit-learn

Frontend:

  • EmberJS

External Services:

  • Akismet (Spam Filtering)
  • api.ai (NLP)
  • IBM Watson API (speech synthesis)

Architecture Diagram:

Forli - Techstack

Video: https://www.youtube.com/watch?v=7R2-YFpCaHg

Presentation: https://goo.gl/1rQKTk

Mock: Forli - Home

Attached few mock screens of various components.

Clone this wiki locally