BRIAN
MUTAI
There's a moment I build for , when something we sketched on a whiteboard is live, and real people are using it. I'm Brian, a full-stack engineer who thinks in products, not just code.
I obsess over the problem until the solution is obvious.
The moment when something we drafted on a whiteboard is running in production and real people are using it. That moment is why I build. I came into engineering through a business lens: I think about problems before I think about solutions, and I care about outcomes more than output.
Beyond client and team work, I'm building my own SaaS , because I believe the best way to understand what founders need is to be one. Having a business background means I sit comfortably at the intersection of product, engineering, and market , and I bring that perspective to everything I work on.
Outside the keyboard: you'll find me on a football pitch, in the gym, or behind a camera. I think staying physically sharp and curious outside of work makes you better at everything inside it. Health isn't a side thing , it's the foundation.
WHAT I DO
- Build products end to end, from first sketch to production
- Integrate AI that's useful, not just impressive
- Partner with founders as a technical co-builder
- Architect systems designed to grow with your product
Core Expertise
- Full-Stack Product Engineering
- AI Features & Integration
- DevOps & Fast Shipping
- Systems Designed to Scale
MOGADISHU · SOMALIA
SELECTED
WORKS
001 / 002
AI-Powered Recruiting Platform
Transparent AI-powered candidate matching with a live Think → Act → Act → Speak MCP workflow. Natural language queries, real-time decision timelines, and explainable ranking. Built with Next.js, AI SDK, and Google Gemini.
View Deep DiveSomalia LMIS
National livestock market information system built for the Somali Ministry of Livestock, Forestry & Range. Real-time price tracking, market analytics, and role-based data access across 16 regions and 33+ active markets, deployed on a government domain.
View Deep Dive
AI Fluency
LLM Integration
Vercel AI SDK for streaming completions and structured output; Google Gemini and Claude for production feature work; Zod schemas to keep LLM responses deterministic and safe.
Agent Orchestration
Custom MCP servers connecting AI to live data sources; multi-step agent workflows with real-time streaming visibility at every stage, from planning through execution.
AI-Native Development
Claude Code and Cursor for accelerated engineering; Lovable and v0 for rapid UI prototyping; AI-assisted code review, documentation, and architecture decisions.
Explainable AI Features
Building AI that shows its work: streaming decision timelines, confidence signals, and user-facing reasoning. Users trust results more when they can see how they were produced.
Capabilities
FRONTEND
BACKEND
DEVOPS & CLOUD
DATABASES
TOOLS