Job Description
I need help putting together a user friendly multi-agent environment set-up for my own personal use to do: a) agentic programming to build software, project development and deployment, and b) research. I’ll almost exclusively be working on software development, tooling, and apps.
There is a ton of information online from people using different multi-agent workflows, IDEs, and platforms and it is a bit confusing figuring out what to use. Let me know if you have an account on x.com so I can share links to some posts of ideas that look interesting.
I want to work with someone who has extensive, daily experience using agents for multi-orchestrated project development. This person should also have extensively tested different agents and automation tools:
- Running multi-agent teams in Claude Code successfully, using tools like ‘Superpowers’ and ‘Skills’
- Using multiple LLM’s in a workflow so not to depend on one model (Opus, Gemini, GPT collaborating)
- Have tried Openclaw (and variations of it like Ironclaw that provide greater security) that are fully-open where there is more customization and potential to run locally; also platforms like Perplexity Computer that are closed but easy to use
I want to run the set-up on a VM so that I have full control of the server and its private, I have access to the top GPUs, tasks can run 24/7, resources can scale if/when necessary, I can get notifications on my phone, and I could move everything local if I wanted.
Here is an example of a work flow that looks interesting to me:
Real-time search → Grok 4.20
Planning → Opus 4.6
Coding (complex) → Claude Code (Opus 4.6)
Coding (well-defined) → GPT-5.3 Codex X High
Write Tests → Gemini 3.1 Pro
Run Tests → GPT-5.3 Codex
Debug → Opus 4.6 (1M)
Some technical guardrails required:
1. Implement a checkpointing system for long-running swarms (orchestrator) to ensure no memory is lost during multi-hour processes.
2. Token loop prevention: A credit monitor to prevent recursive loops from exceeding a budget.
Required Skills:
Multi-Agent Orchestration: Proven experience setting up Claude Code with custom tools and skills.
AI Model Integration: Expertise in configuring multi-model workflows involving Anthropic (Opus), Google (Gemini), and OpenAI (GPT).
Linux Server Administration: Advanced knowledge of VM management, driver installation (CUDA/NVIDIA), and server-side security.
Open Source Agent Frameworks: Daily experience with open-source tools like Openclaw, Ironclaw, or similar customizable platforms.
Network Security: Ability to configure firewalls and egress filtering to prevent unauthorized data transfer (data exfiltration).
Python/Scripting: Proficiency in writing the automation scripts necessary to connect different agents and manage task handoffs.
Apply tot his job
Apply To this Job