SPV and MDO be able to reach out to their 30M userbase for targeted marketing and feedback mechanism.
Feature Value
SPV/MDO will be able to connect with users, understand their issues, nudge them to enroll, complete course consumption. With the data insights, SPV/MDO be able to take decisions how to promote learning on iGOT.
Feature Vision
SPV/MDO be able to create user buckets based on certain criteria. E.g. users who have enrolled in at least 1 course etc.
SPV/MDO be able to query and create user buckets by querying in natural language (NLP and LLM). e.g., A SPV may say, “Create a list of users who have been active for at least 5 hours in past 24 hours”.
Upload SMS, WhatsApp Messages formats and email pre-defined templates.
SPV/MDO be able to enable feedback mechanism and enabling analytics and AI driven insights from the responses.
Goals & Mid-Point Milestone
Develop a Communications Console integrated with iGOT Platform, enabling SPVs and MDOs to engage with a user base of 30 million through SMS, WhatsApp, and Email channels.
Facilitate targeted marketing campaigns and feedback mechanisms to enhance user enrollment and course completion rates.
Provide data insights to SPVs and MDOs for informed decision-making on promoting learning initiatives on iGOT.
Expected Outcome
Improved user engagement and communication channels between SPVs, MDOs, and the iGOT user base.
Increased user enrollment and course consumption through targeted marketing campaigns and personalized communication.
Enhanced data analytics capabilities to derive actionable insights from user feedback and interactions.
Acceptance Criteria
Communications Console successfully integrated with iGOT Platform, allowing SPVs and MDOs to access SMS, WhatsApp, and Email channels.
Ability to create user buckets based on specific criteria such as course enrollment status, activity level, etc.
Natural language query capability implemented using NLP and LLM for creating user buckets.
Upload functionality for SMS, WhatsApp messages formats, and email pre-defined templates.
Feedback mechanism enabled with analytics and AI-driven insights derived from user responses.
User feedback indicating satisfaction with communication effectiveness and responsiveness.
Implementation Details
This project has to be built from scratch and is tech stack agnostic
Making a prod ready authentication sytem using nextJS , expressJS and PostgresDB
Integrating s3 into no code email editor
Learning zustand to make the entire no code email editor work out of the box
Week 2
SMTP , SMPP , etc . The types of protocols used for different media communication and their workings
Tried setting up SMTP server on my own
Search for a configurable SMPP server that could be hosted / be used for free to develop
Researched on Whatsapp Bussiness APIs
Week 3
Learnt about designing a better LLM chat interface
Learnt to use langchain and prompting techniques to make the accuracy better
Learnt the importance of sanitization
Researched on how to use loggers for logging , user query and LLM converted DB query . This will help us to create a dataset for our own SQL LLM.
Week 4
Learnt about nodemailer and tested things out
Learnt about various headers in emails like ReplyTo , etc
Learnt about X- headers in emails , that can be used to create custom headers
Learnt about DKIM signing
Learnt how to use bullMQ and Redis .
Learnt about redis pooling and callback
Witnessed the practical working of consumer and workers
Learnt about Redis complex data structures
Picked up on how to delay redis queueing
Week 5
Learnt how does company tracks us , and how can they check if we have seen the mail or not , Mail pixels are the answer to this
Learnt about various HTTPS headers that can be received when the API is hit , including device type and location
Planned out prisma schema changes according to the research done above
Week 6 and Week 7
Setting up channels UI using shadcn-chat components
Learnt about how mail threading works in mail client
researched on common headers that spans across multiple email clients to create custom threading solutions to prevent breaking of threads on our DB
Played around with X- headers and various headers to make sure all emails are caught correctly
Learnt about clients changing reply ID , and break threadID , so develop conditions to prevent breaking in our threads
Week 8
Learnt about nginx , pm2 and other best practices to deploy backend
Learnt about .env handing by ENV_NODE , and how it can make it easier to handle envs on prod and development level
Week 9
N/A
Week 10
Learnt about how we can cut down dockerfile size by a lot by using nginx image combining with node and multistage build . Ever more lighter than node alpine !
Learnt on maintain configs in much better way by hit and trial method
Ticket Contents
Communications Console
SPV and MDO be able to reach out to their 30M userbase for targeted marketing and feedback mechanism.
Feature Value
SPV/MDO will be able to connect with users, understand their issues, nudge them to enroll, complete course consumption. With the data insights, SPV/MDO be able to take decisions how to promote learning on iGOT.
Feature Vision
Goals & Mid-Point Milestone
Expected Outcome
Acceptance Criteria
Implementation Details
This project has to be built from scratch and is tech stack agnostic
Product Name
Karmayogi
Organisation Name
MeitY
Domain
Learning & Development
Tech Skills Needed
Other
Mentor(s)
@mbcse
Category
Backend