AI Automation Engineer · Nairobi, Kenya
I architect agentic AI systems and deploy intelligent automations that eliminate the manual work draining your time, your team, and your revenue. Built with Python, FastAPI, n8n, Make.com, GoHighLevel, and the latest LLMs.
Tools & Platforms
About
There's a version of your business where the repetitive work just happens — leads get followed up, data gets logged, reports get generated, decisions get made — while you focus on the work that actually needs you. That's what I build.
I'm Joshua Onchaga, an Agentic AI and Automation Engineer based in Nairobi, Kenya. I specialise in designing multi-agent systems and end-to-end automation that connect your tools, inject AI where it adds real leverage, and eliminate the manual tasks that consume your team's time without producing meaningful output.
I got into agentic AI because I'm wired to see inefficiency and immediately want to fix it. Watching a business owner spend three hours on tasks a well-built agent pipeline handles in three minutes — that bothers me deeply. Building systems that give those hours back is genuinely fulfilling work.
I've worked with agencies, SMEs, coaches, and consultants across sales, marketing, operations, and customer support — delivering systems that don't just run, but keep running, with clear documentation and zero hand-holding required.
Sample Work
01 / 04
The Problem
Appointment booking was handled manually staff fielding calls, checking calendars by hand, sending confirmations one by one. Timezone mismatches caused missed slots and staff time was consumed by a task that needed no human judgment.
02 / 04
The Problem
A content team was spending hours manually writing platform-specific posts, generating images, and copy-pasting to LinkedIn, Instagram, and Facebook. The process was slow, inconsistent, and bottlenecked on one person.
03 / 04
The Problem
An errand service was logging orders by hand, sending status updates individually, and losing track of which orders were assigned, in-progress, or completed. Customer experience suffered while staff spent more time on admin than delivery.
04 / 04
The Problem
A business needed high-volume content across multiple formats and channels simultaneously requiring multiple tools, multiple people, and hours of coordination. Output was inconsistent and the pipeline fragile.
What I Do
01
Multi-agent pipelines where specialized agents coordinate, hand off tasks, and deliver results. Built with Python, FastAPI, and LLMs. Research → Analysis → Synthesis, end to end.
02
End-to-end automation of repetitive business processes using n8n, Make.com, and Zapier. If it's manual and recurring, it can be automated.
03
Embedding LLMs (Gemini, GPT-4, Claude) into workflows for intelligent classification, content generation, decision routing, and autonomous reasoning.
04
Connecting CRMs, databases, spreadsheets, and external APIs via REST, webhooks, and FastAPI. Clean, documented, production-ready.
05
Discovery sessions to map current manual workflows, identify the highest-leverage automation opportunities, and produce a clear blueprint before any tool is touched.
06
One-on-one and small group training for teams who want to understand, build, or manage their own automation and agentic AI systems — practical and jargon-free.
If your team is doing the same tasks on repeat, the solution already exists. Let's talk about what automation can do for your business — no jargon, no pressure, just a practical conversation.
Current Status