I will build claude code agents and multi agent ai systems for your company


Vetted Pro
Top Rated
Vetted by Fiverr Pro
Joshua D was selected by the Fiverr Pro team for their expertise.
About this gig
Most teams use Claude like a chatbot. Set up properly, it works like a team of agents that plan, build, test and review work, and it costs a lot less to run.
I run Ten Peaks Tech, a small software agency, and we run our own work this way. One lead agent plans and makes the hard calls, cheaper specialist agents handle searching, testing and review, and hooks and budgets keep token spend under control.
What I can set up for you
- Claude Code for your repo with CLAUDE.md, skills, subagents and hooks
- Multi-agent architecture with model routing to cut token costs
- MCP connections to GitHub, Jira, Notion, databases and internal APIs
- Custom agents on the Claude Agent SDK for support, reporting, ops or sales
- Guardrails, logging and cost tracking so you can see what agents are doing
You'll deal with me directly, on calls and in messages, not a bot or an account manager. Message me with what you want automated and I'll tell you honestly whether agents are the right fit.
Get to know Joshua D
Vetted Pro
Making your Application Ideas a Reality
Top Rated
Joshua D is part of the Fiverr Pro catalog and has been hand-picked by a dedicated Fiverr Pro team for their skills and expertise.
Vetted for
Software Development
- FromUnited Kingdom
- Member sinceJul 2022
- Avg. response time2 hours
- Last delivery4 days
Languages
English
My Portfolio
Other Software Development Services I Offer
FAQ
What is a multi-layered agent setup?
One lead agent on the most capable model plans the work and makes the decisions. It hands searching, testing and routine coding to specialist agents on cheaper, faster models, then checks what comes back. You get top-model quality where it matters at a fraction of the cost.
How much will it save on tokens?
It depends on the work, but moving reading, searching and test runs to smaller models makes a big difference. I set up usage tracking so you can see the numbers yourself.
Can the agents use our tools?
Yes. Through MCP servers or custom tools they can work with GitHub, Jira, Notion, Google Drive, your database and internal APIs.
Is our code and data safe?
Permissions and hooks control what agents can read, run and change, and anything risky needs a person to approve it.
Do we need a Claude plan?
Claude Code needs a Pro, Max or Team plan per developer. Custom agents run on an Anthropic API key in your account.
