1
1
Kenya
2+ years of experience
I build full-stack AI-powered products. Define the problem → design the system → orchestrate AI to build it → ship. Live: SkillOS Lite (flyrankaiprojects.vercel.app) — streaming AI skill-evaluator, FlyRank capstone. Stack: Next.js 16, TypeScript strict, Tailwind v4, Vercel AI SDK, Gemini. In development: CareConnect (healthcare SaaS), ACCORD (humanitarian crisis response), DesignOS (AI design studio). Hackathon: Aegis Agent — Band of Agents Hackathon, lablab.ai. Multi-agent investment analysis. Certificate earned. Open source: SkillOS co-creator — the open standard for AI agent behaviour (MIT licensed). Real engineering foundations: HTML/CSS solid. JavaScript building. AI orchestration, prompt engineering, and full-stack product thinking are genuine strengths. Open to hackathons, collabs, and first paid roles.

Enterprise investment due diligence is slow, opaque, and inconsistent. Analysts spend hours collecting information across disconnected tools with no systematic review process and no audit trail. Aegis Agent solves this with a genuine multi-agent workflow. Four specialized AI agents — Forensic Planner, Data Analyst, Risk Auditor, and Executive Finalizer — collaborate through Band chat rooms to analyze investment opportunities. Every agent handoff happens via Band. Every decision is logged. When a user submits a company and deal context, Aegis creates a dedicated Band room and the agents execute a structured workflow: 1. Planner (AI/ML API): Decomposes the task, selects optimal inference models 2. Analyst (Featherless AI): Synthesizes findings, cites sources with URLs 3. Reviewer (Featherless AI): Validates analysis, scores risk 0–100, loops back if gaps found 4. Finalizer (AI/ML API): Produces polished dossier with verdict, alternatives, and reasoning chain Users can pause the room and intervene at any point — the agents adapt and continue with user direction baked into their context. The frontend renders this entire operation as a live war room dashboard. Acts are cinematic. Handoffs are visible. Partner tool usage is unmissable. The result: auditable, explainable, enterprise-grade AI investment analysis.
19 Jun 2026