Álvaro
Data Scientist
“In a climate this uncertain, Shakers gives you a layer of reassurance and confidence. What I like most, without a doubt, is the flexibility and getting to work on top-tier projects.”
Join ShakersShakers for expert talent
Work as an AI Agent Developer
Projects where the agent is yours: which tools it touches, what memory it has and what it refuses to do. Shakers connects you with companies that want agents working, not demos.
The trade reads in two places: the tutorial that assembles an agent in 20 minutes and the prediction that agents will do everything alone. Neither tells how a real company's agent is sustained, engagement by engagement. Here is that.
- Higher-paying projects
- More continuity between projects
- Greater recognition for your experience
+3,000
Tech projects published on Shakers
+90
Tech, Product and Data specialisms are part of Shakers
6 months
Average length of a project with Shakers
4.7/5
Our AI Builders' satisfaction on Trustpilot
AI Builder careers powered by Shakers
See how other AI Builders access better projects, build their reputation and grow their careers.
Amelia
AI Agent Developer
“I have been building AI agents for years. Thanks to Shakers I stopped chasing clients and started genuinely choosing which projects I wanted to work on. I work with some of the best companies in Europe, at my own rate, and I focus on what I am good at and what I love: building.”
Join Shakers
Luz
Backend Developer & Data Engineer
“Shakers adds a lot of value because it connects you with sharper projects, already filtered and better defined. It saves you a big part of the upfront work of understanding the client, quoting and weighing up whether it is worth it, and that means that as a contractor you can go far more directly to the opportunities that genuinely fit you.”
Join Shakers
Rubén
Software Developer
“Life as a contractor has a lot going on and it is not always easy. We like to focus on what we are good at, but you also have to manage clients, invoicing and go looking for work. For me, Shakers has been key because it does that prospecting for me, it gives the client confidence from the very first moment, and it lets us be part of a community with real enthusiasm.”
Join Shakers
Alejandro
AI Developer
“Shakers was key to getting my company off the ground: it gave me the flexibility to keep contracting while I built my next chapter. Without Shakers, I would have had to raise funding or push the project back.”
Join Shakers
Toño
Product Designer
“With Shakers I have landed recurring projects that give me peace of mind and room to grow professionally. It is not just about reaching quality clients: it is feeling that you have a network behind you, one that connects you with real opportunities and helps you move forward as a freelancer.”
Join ShakersWhat an AI Agent Developer decides every week
An AI Agent Developer builds agents that execute business work: orchestrating the tools the agent calls, designing its memory and context, chaining agents when the process demands it, and adding the guardrails that prevent incidents. They don't train the model (AI Engineer) or write the master prompt (prompt engineer): their trade is the agent that reasons over a process and acts.
You decide which tools the agent calls and with what permissions, what it remembers and what it forgets, where its decision ends and human approval begins. You also decide what the agent refuses to be: without evaluation or limits there is no production. The process belongs to the client; the agent that sustains it, to you.
PythonLangChainLangGraphMCPfunction callingRAGmemorymulti-agentguardrailsagent evaluationOpenAI APIAnthropic APIvector DB
OrchestrationMemoryGuardrailsMulti-agentMCPEvaluation
Free to join, no hidden fees
How Shakers works for talent
Create your profile
Tell us who you are, what you can do, the projects you want to work on and what your availability and rate are.
Get your certification
We certify your experience and your use of AI so companies understand and trust what you bring.
Match with projects
We connect you with well-paid projects that fit your expertise and your preferences.
New Roles for AI Era projects that live up to your expectations
Work on real projects, with teams that need your expertise, without wasting time looking for opportunities.
Projects that fit you
Work with companies that need your stack and experience to take their projects to production.
Opportunities that come to you
Our AI matching connects your profile with the projects that fit your preferences.
You build. Shakers handles the rest
Set your own terms based on your seniority. We manage contracts, payments and paperwork.
What an AI Agent Developer delivers on a Shakers project
The brief arrives with the process
Shakers is hiring infrastructure: the project arrives with the process the company wants to turn into agents and its tools, not a list of technologies.
An agent you can audit
You deliver the agent, its guardrails and the log of what it did, including the case you dropped and why it doesn't automate.
What happens when the process changes
You set what gets re-evaluated when a tool or the process changes, and who answers for the agent the day it hesitates.
Agents sustained, not demonstrated
Having put agents in production and being able to tell which guardrail saved you, and what you did when the agent decided badly.
Agent, flow or nothing
Recognising when the process asks for an agent and when simple automation does: the honest answer beats selling agents.
How far your brief goes
You orchestrate the agent; you don't train the model or design all the context. If the project needs that, you say so at kickoff.
Frequently asked questions
Any other questions? Write to us and we will reply.
What does an AI agent actually do?
An agent receives a goal, decides the steps with the model and acts on your tools: reads, writes, queries and executes. Versus a flow: it handles the unexpected, not just the programmed.
How do you become an AI agent developer?
If the question is building them: Python, LangChain and real processes. What the certification measures is agents sustained in production, not tutorial builds.
What tools does an agent developer use?
LangChain and LangGraph to orchestrate, MCP and function calling to connect tools, and OpenAI's or Anthropic's APIs as the model. The judgement: which agent each process deserves.
How much does a freelance AI agent developer earn?
You invoice per project, and scope sets the band: tools to connect, memory and required guardrails. The budget is agreed before accepting.
Can AI agent projects be done fully remotely?
Almost all: agents run in the cloud and sessions with the team happen remotely. Only the occasional kickoff asks for presence.
How do I join Shakers as an agent developer?
Create your profile, pass the certification process and join the collective: only 4 percent make it. Then you choose which agent projects to work on.
The day to day of the trade
The connector that breaks when the API changes, the agent that works in the demo and improvises on Friday, the memory that grows without criteria. The trade is learned with agents in production.
The market asks for it embedded
3,352 live UK tech vacancies mention ai agent, according to Shakers' market analysis (n=39,811, September 2026), inside AI Engineer titles. Whoever already builds agents finds projects in the collective.
From tutorial to engagement
The frequent door is the maker who automated their own work with agents. The leap comes when the agent belongs to someone else, with guardrails and someone waiting for the result.
Identity finally has a name
AI Engineer, prompt engineer, agent developer: the SERP confuses them. The frontier: you orchestrate the agent; the AI Engineer serves the model; the prompt engineer sets its behaviour.
Roles that cross an agent project
Four roles from the collective cross your path on an agent project, from who serves the model to who sets its behaviour.
AI Engineer
Builds and integrates the model your agent reasons over: the capability the agent developer orchestrates into work.
Prompt Engineer
Sets the behaviour of the model your agent uses at each step: versioned, evaluated prompts inside the flow.
Context Engineer
Designs the data your agent receives: RAG, memory and curation. Without context, the best agent improvises.
Python Expert
The custom integrations the agent needs when standard tools don't reach: from connector to deployment script.
The boundaries between these roles are agreed on each project: the title does not set them.
Take the next step with Shakers
Want projects where the agent is yours?
Pass Shakers' certification and choose your agent projects.
