Projects

Working systems, hard lessons, and builds I chose to throw away.

This is the deeper layer of the site. The front door is intentionally simpler now, but the evidence stays here: what problem was real, what I tried to build, what the architecture taught, and what still needs work.

If you want the fastest read on the current work, start with Founder OS, ATOM + Joan, and Sidekick. Everything else is supporting evidence, an older build, or advisory context.

Start Here

Start here

The three things that matter most right now: the system running my business, the method behind the work, and the first customer-facing product family.

  1. Founder OS

    Primary

    Founder OS briefs me, researches accounts, plans work, prepares follow-up, and keeps review points around anything that can create risk.

    Founder OS · Running BusinessDaily briefs | Review gates | Real business work
    Case Study

  2. Joan

    Primary

    Joan is the direction for pulling that context into one product layer: what happened, what changed, who reviewed it, and what should happen next.

    ATOM + Joan · Method + CompanyShared context | Review trail | Product direction
    Case Studyjoanai.io

  3. Sidekick

    Primary

    Sidekick folds the old meeting, sales, and team-sidekick ideas into one product family: prep the meeting, capture what matters, close it out, and carry the work forward.

    Sidekick · Meetings + RevenueMeetings | Revenue work | Follow-through
    Case StudySalesSidekickSidekick Genius

Supporting Evidence

Supporting evidence

Supporting systems that show parts of the same thesis: engineering discipline, client delivery, meeting workflows, wealth-management work, signal capture, and the site itself.

  1. VibeOS

    The development system behind the rest.

    HarnessTwo years of fixing what breaks and adapting to model advances.
    Case Study

  2. Meeting Sidekick

    A digital employee for every meeting.

    Meeting lifecyclePrep, in-call capture, and a 6-output closeout.
    Case Study

  3. UYSP

    An AI BDR that works the top of the funnel.

    AI BDR60+ workflows plus a 73-route, 19-table CRM.
    Case Study

  4. Joans

    Wealth-management AI for fiduciary client work.

    Wealth managementMicrosoft foundation, secure client journey, and service controls.
    Case Study

  5. SignalClaw

    Capture and memory for AI-assisted work.

    Active capture40 migrations, about 1,600 tests, 9 architecture rules.
    Case Study

  6. Chief Latif Website

    An interactive site for Latif's work, projects, AI guide, and fit analysis.

    Interactive personal siteProject-aware AI guide, fit analysis, knowledge graph, and 399 backend tests.
    Case Study

  7. Joan-specialists

    The specialist-worker incubator on top of Joan.

    JoanSpecialist-worker incubator; cognition pending.
    Case Study

Origins and Archive

Origin stories and archived builds

Earlier systems and parked builds that still matter because they explain why the current architecture is simpler.

  1. SalesSidekick Legacy

    The first monster build, finished and deliberately thrown away.

    Origin storyBuilt four times, about 5,000 tests, 50-user beta GO.
    Case Study

  2. WorkspaceOS

    Cowork, built a year early in Cursor, then thrown away.

    Right earlyCowork-like primitives built about a year before Anthropic shipped Cowork.
    Case Study

  3. Fabriq

    AI capital intelligence connecting spend to workforce outcomes.

    AI spendPrivacy-preserving AI capital intelligence, now part of the Joan history.
    Case Study

  4. Work Intelligence Platform

    Revenue intelligence plus proactive work routing.

    Work intelligenceCode-mature non-Joan candidate, parked as earlier project history.
    Case Study

Advisory and Community

Advisory and community

Startup advisory, founder education, and public artifacts that support the work without becoming more brands on the front page.

  1. Plug & Play Advisory

    AI GTM mentorship for Plug & Play startups.

    Advisory70+ startups mentored through the accelerator.
    Case Study

  2. Skool Community

    A free community for AI sales and GTM.

    CommunityWeekly networking, office hours, and training materials.
    Case Study

Community Projects

Public repos and shared frameworks

These are the public artifacts I share more openly: tools, frameworks, and workflow libraries that extend the portfolio beyond the core case studies.

  1. VibeOS Bootstrap

    Open Source

    Open-source setup for adding quality checks to an AI development project.

    Public frameworkProject bootstrap, quality checks, decision trees, and agent setup for Claude Code, Cursor, and Codex.
    GitHub

  2. Reddit Scout

    Open Source

    Reddit discovery and analysis tool with a searchable dashboard and sentiment views.

    Public frameworkCommunity intelligence, discovery workflows, and lightweight product experimentation.
    GitHub

  3. Prompt Engineering Bible

    Open Source

    A framework for managing AI prompts across complex systems.

    Public frameworkPrompt contracts, safety rules, quality standards, and evaluation criteria for AI applications.
    GitHub

  4. Google MCP Server

    Open Source

    Connects AI agents to Google Workspace - Gmail, Calendar, Docs, Drive, and Sheets.

    Public framework28 tools with authentication, error handling, and production-ready reliability for AI integrations.
    GitHub

  5. SalesSidekick n8n Workflows

    Open Source

    Open-source automation workflows from the SalesSidekick and client delivery systems.

    Public framework50+ workflows for CRM sync, data processing, monitoring, backups, and sales operations.
    GitHub