Skip to content

AI Agentic Developer / Nairobi, Kenya

I BUILD SOFTWARETHAT CAN THINK,ACT & EXECUTE.

AI agents, intelligent systems and full-stack products engineered for the real world.

Software was the beginning.
Intelligence is the next layer.

I am a Kenyan software engineer focused on building AI-powered products and agentic systems that turn complex workflows into intelligent, executable software.

Based In
Kenya
Focus
AI Agents / AI Systems
Stack
Python / LangChain / LangGraph / n8n / TypeScript / React / Next.js
Building
Intelligent Software

AI Engineering Stack

The tools I use to design, orchestrate and deploy intelligent systems.

AI Agent Frameworks

AI Automation

AI Engineering

Application Engineering

The Engineer Behind The Systems

Location
Kenya
Focus
AI / Software Engineering
Role
Builder / Researcher / Entrepreneur

I’m James Thuranira, a Kenyan software engineer building at the intersection of artificial intelligence, software engineering and real-world problem solving.

My work spans AI agents, agentic workflows, full-stack applications, machine learning concepts and experimental technology products.

I am particularly interested in what happens when software stops being passive and starts becoming capable of reasoning, deciding and acting.

What I Build

  • 01AI Agents
  • 02Agentic Workflows
  • 03Intelligent Automation
  • 04AI Products
  • 05AI Engineering Systems
  • 06Full-Stack Applications
  • 07Machine Learning Systems
  • 08Data Platforms
  • 09Developer Tools

Selected Systems

Things I’ve built, shipped and experimented with.

AI Agent Systems

From prompts to systems.

An AI agent is more than a model call. I design systems where models can reason over context, use tools, interact with APIs, maintain state and execute multi-step workflows.

Agent

The orchestrating layer - decides what needs to happen and in what order, using LangGraph to manage state across steps.

Engineering This Requires

  • LangChain for model and tool integration
  • LangGraph for stateful agent orchestration
  • Python for AI services and data processing
  • n8n for workflow automation and external integrations
  • Zapier for business application automation
  • APIs and databases for real-world execution

Capabilities

What I take from idea to production.

01

AI Agents

Autonomous systems capable of reasoning, tool use and multi-step execution using frameworks such as LangChain and LangGraph.

02

AI Products

End-to-end applications powered by modern AI models, structured workflows and intelligent interfaces.

03

AI Automation

Turn repetitive business workflows into intelligent systems using n8n, Zapier, APIs and custom Python services.

04

Full-Stack Engineering

Production applications from interface to database and infrastructure.

05

Machine Learning

Custom predictive, recommendation, classification and optimization systems using Python and modern machine-learning tools.

06

AI Integration

Connect AI models with APIs, databases, CRMs, internal systems and external tools.

I don’t just add AI.
I design systems around it.

01

Intelligence

AI should understand context rather than simply generate text.

02

Execution

Useful AI should be able to interact with the systems around it.

03

Impact

The technology must solve an actual problem.

The Lab

Ideas I’m exploring before they become products.

AI agent architectures
AGIOS - an operating layer for agents
Custom ML systems
Autonomous workflows
AI + marketplaces
Intelligent dashboards
Agent memory systems
Human-agent collaboration
LangGraph agent orchestration
n8n-powered automation
Zapier business automations
AWS-based AI infrastructure
Multi-agent systems
AI evaluation & observability

Built in Kenya.
Designed for the world.

I build from Nairobi, working across the same problems that shape emerging markets - mobile-first infrastructure, automation and financial systems - while engineering software meant to hold up anywhere.

02+AI / software products
10+Technologies used
01Independent builder
Ideas being explored

Have a problem
worth building?

Let’s turn it into an intelligent system.