Product Roadmap

Type

Web (Node) (Build-On)

Overview

Organization: Underdog Devs

• Mission: Help formerly incarcerated or disadvantaged people find careers in tech
• Founded by BloomTech grad and many BloomTech grads involved
• Mostly volunteers
• Mentor/mentee structure

Notes:

Need a platform to manage people and resources and track operations**:**

• Recommend individuals for job opportunities, stipend/grant opportunities
• Intake process for members
• Administrators assign mentors to mentees
  • Manual process, but administrators receive recommendations from the platform on how to make matches
  • Mentors and mentees schedule their own meeting times
    • Scheduling happens between mentor and mentee via their own informal methods (e.g. Slack), want to keep it this way to avoid imposing additional complexity and prioritize the relationship
    • Mentors and mentees can optionally use a built-in calendar to schedule and track sessions
      • Need a place for the mentor to put the meeting link (e.g. Zoom link, phone number) (could probably just be a text field)
      • Alternative: Mentors could use a “check-in” button to use at the start of a meeting to track meeting times in a low-friction
  • Right now the mentor reports back whether their mentee is not responsive or engaged
    • Platform should let mentors report when meetings were scheduled, whether it actually occurred, any “memos” coming out of the meeting
  • Need to track who’s receiving internal resources (physical equipment, books, slots in educational programs)
    • Some tangible, some service-based (even internship opportunities)

The organization is currently trying to organize using a Trello board:

• Intake of potential mentor/mentee matching
• Separate spreadsheet for tracking actual matches
• Slack community that all members are added to
• Prepare resources for people—job search, interviewing, etc.

Organizational partnerships involved as well:

• Hiring partnerships
• Offering mentorship
• Providing resources (equipment, education)

Contacts

Jason Long: [email protected]

Technology Stack

• Web BE: Node
• Web FE: React
• Auth: Okta → Moving to Auth0
• Storage: MongoDB

User Base

• Superadministrator
  • Manage users & roles
  • All Administrator permissions
• Administrator
  • Interview process:
    • Vet new member application (Mentors & Mentees)
    • Scheduling application interviews
    • Sending interview comms via the platform (Sendgrid)
  • Onboard new members (Mentors & Mentees)
    • Discipline-specific (e.g. mobile dev, web dev)
    • Get assessment of their skill/experience level
      • This is currently live interviews → standardize into quizzes (automated assessments)
      • Further explore with stakeholder—3p code challenge scores in some cases?
  • Assign and track who’s receiving internal resources (physical equipment, books, slots in educational programs)
    • (Mentees are always recipients of resources)
    • Dashboard
    • Probably most useful to break these down visually by resource type
    • Track monetary value for tangible resources
  • Track operations
    • Coordinate what programs they have (some via partnerships)
    • Track contacts & applicants related to those programs
    • Tracking mentor/mentee meetings
      • When meetings are for specific individuals involved
      • Whether meetings happened
      • Notes related to meeting
      • Escalations attached to meeting? (e.g. conduct)
    • Track "tickets" (e.g. not responsive, conduct) connected to escalations (”action tickets”)
  • Track recommendations made by mentors for mentees to receive resources (or applications made by mentees to get resources)
    • Track status of each recommendation as a ticket ("Ready for Review" → "Under Consideration" → "Approved" / "Rejected")
    • Track who made the request (mentor or mentee)
    • If approved, administrator coordinates provision of resources and updates the ticket accordingly
      • The ticket triggers an action to adjust inventory and resource allocation
• Mentors
  • Apply to be a mentor via the platform
  • Keep track of who assigned mentees are
  • Track mentees' current education progress
  • Report when meetings were scheduled, whether it actually occurred, any notes coming out of the meeting
  • Report back whether their mentee is not responsive or engaged
  • Recommend mentees for resources, job opportunities, stipend/grant opportunities
• Mentees
  • Apply to be a mentee via the platform
  • Receive any application process communications via email (Sendgrid)
  • Identify what their expectations are
  • Access resources from a centralized location
    • Equipment
    • Programs
    • Informational
  • Apply to get access to resources
  • Meet with mentors

Feature Name	Description	Release
User Roles	Implement all user roles using Okta Auth0.	Release 1
User Management	Administrators should be able to manage users in Okta Auth0, including: - Creating new administrators - Creating and deleting users	Release 1
Auth0	We’ll need to refactor the codebase to accommodate Auth0 instead of Okta.	Release 1
Application Submission	Mentors and Mentees should be able to submit applications to Underdog Devs.	Release 2
Ticket System	Administrators and Mentors should be able to view "tickets" with a status and additional fields.	Release 2
Two-Factor Authentication	Implement two-factor authentication in Auth0.	Release 2
Application Tickets	Administrators should be able to view tickets created automatically for each new application to review, vet, and approve/deny each.	Release 3
Mentor/Mentee Matching	Administrators should be able to assign mentees to mentors based on prior application and availability data (as available). Work closely with the Data Science Manager and the stakeholder to formulate questions and a model.	Release 3
Visualizations	Administrators should be able to visualize real-time information relevant to them, such as: - How well are available mentors distributed among mentees, both in general and for specific needs? Work closely with the Data Science Manager and the stakeholder.	Release 4
Model Interpretation	Provide administrators with plain-language insights into why the ML model made the decisions it made. Work closely with the Data Science Manager and the stakeholder.	Release 4
Meeting Tickets	Administrators should be able to view tickets created for each meeting and log relevant events and information. Mentors should be presented with a meeting ticket after meeting with a mentee where they give a writeup of what was covered. These writeups should be aggregated into summaries and categories for: - all meetings for an individual mentor/mentee pair - all meetings Work closely with the Data Science Manager and the stakeholder.	Release 4
Resource Tickets	Administrators and Mentors should be able to create and view tickets for resource recommendations as well as deployed resources. Both user groups should be able to log relevant events and information	Release 5
Escalation Tickets	Administrators should be able to create and view tickets for escalations. Work closely with the stakeholder.	Release 6
Email Communications	All user groups should receive communications relevant to them via email.	Release 7
Alerts	Administrators have alerts displayed to them featuring things requiring action/attention.	Release 7

Potential Data Science Features

• Application Review
  • Very few cases in which applications are denied
• Mentor/Mentee Matching
  • Understanding domain spaces: NLP to standardize/categorize text input
• Recommendations for job opportunities
  • NLP on job descriptions vs resumes?
    • K-nearest neighbors model
    • Similarity between resume and job description
    • Whether those who applied to similar jobs got the job
    • Whether/how long they kept the job
    • Need web functionality to gather this data
  • “What is the most meaningful factor/point of contact that results in a candidate landing a job?”
    • Every mentee is paired with one or more mentors
    • Need to track engagement with mentors
      • Anecdotally, the most successful hired mentees never missed a single session with their mentor
      • What % of scheduled engagements are they making?
    • General Underdog Devs community engagement
      • Slack engagement
    • Currently informal resume review sessions
    • Can these data points also inform resource allocation?
    • Data collection from mentors
      • Ticketing data
  • Data will be much more powerful later if it’s labeled data rather than just raw text input—so need to ask specific questions
• NLP to analyze, summarize case notes from mentoring sessions
  • “Mentor, what did you cover in this session?”
    • Predefined checkboxes + open text input
    • Maybe a summary input, but then 2 lists of potential topics to have been covered during a 1:1 (Using that as an example? EX: Technical: JavaScript, HTML/CSS, React, Python, act - Non-Technical: Job Prep, Resume, LinkedIn, Reachouts, Imposter Syndrome/Confidence building, ect.)
• Track trends about what brings people to Underdog Devs
  • Help Underdog Devs recruit mentors that are needed based on what activities/skills are in demand / helpful
  • Ask mentors what they’re teaching, what they intend to teach