| # | JD Requirement | STAR+R Story | S | T | A | R | Reflection |
|---|---|---|---|---|---|---|---|
| 1 | AI strategy + production ML | Kozi Sports Link — built ML prediction pipeline with GenAI layer | Solo founder building production ML system | Need for human-readable AI predictions in sports | Designed ML pipeline + GenAI layer for analysis | 70%+ live prediction accuracy sustained | "I'd approach AI strategy by starting with measurable outcomes, not technology." |
| 2 | Hands-on architecture + code | Fluo ERP — rebuilt from Laravel/Bootstrap to Laravel + React + ML | Self-hosted ERP couldn't scale or support ML | Legacy monolith with no API layer | Refactored to multi-tenant SaaS with ML microservices | 3-product suite extracted from single codebase | "I prototype the architecture myself before asking anyone to follow." |
| 3 | Domain-Driven Design | ERP module boundaries (HR, Finance, Inventory, Reporting) | Growing client needs across different business domains | Single ERP codebase serving all domains | Designed bounded contexts per module with shared infrastructure | Modules operate independently, integrated via APIs | "I design domain boundaries based on business capability, not technical convenience." |
| 4 | Security-first architecture | RBAC + audit logging in ERP for enterprise clients | Enterprise clients required role-based access control | Security was an afterthought in original design | Built RBAC with audit logging, dynamic reporting | Enabled enterprise sales, zero security incidents | "Security is architectural, not a layer you add later." |
| 5 | Distributed systems at scale | Multi-tenant SaaS architecture (Fluo Analytics) | Extracted forecasting engine needed to serve multiple tenants | Monolithic deployment couldn't scale per-tenant | Built standalone multi-tenant SaaS API with isolation | Serves multiple clients from single deployment | "Multi-tenancy requires isolation at the data and compute layer from day one." |
| 6 | Coach and mentor engineers | Mentored 10+ collaborators at Tylique Consolidated | Small agency needed to onboard collaborators quickly | No formal training structure | Created mentorship program around backend frameworks | 10+ collaborators productive; team sustained through COVID | "I lead by making others better, not by being the bottleneck." |
Recommended Case Study: Design a secure, event-driven microservices architecture for a crypto banking platform that needs to support AI-powered features — map to the candidate's ERP-to-SaaS evolution story.
Likely Red-Flag Questions:
- "You don't have 10+ years — why should we consider you?" → "My 6+ years span founder-level ownership of 3 ventures, which is equivalent scope to 10+ years of IC + leadership. I've designed, built, and operated production systems end-to-end."
- "No crypto experience — how will you ramp?" → "I've entered and delivered in unfamiliar domains before (sports analytics, fintech dashboards). The AI strategy and architecture skills transfer directly; crypto domain knowledge I can acquire in weeks."
- "Can you work in CET timezone from Kenya?" → "I'm in EAT (GMT+3), which overlaps 6+ hours with CET. I've maintained async-first delivery across time zones throughout my freelance and founder work."