A freelance AI engineer, hired per project
Four defined offers from one engineer: an AI agent, an MCP server, retrieval over your own documents, or an evaluation of a system you already run. Each is scoped in writing, priced per milestone and handed over complete.
The four offers
AI agent development for one job, done unattended
I build an AI agent that does one defined job inside your business, with the checks, the logs and the handover a person would need to trust it running without them.
- You bring
- A job a person does today, and the systems it touches
- You get back
- An agent in production, with its checks, logs and runbook
- Shape
- Three milestones: one path working, the full job, the handover
MCP server development for your product or your data
I build the MCP server that lets Claude, ChatGPT and other assistants use your product or query your data directly, with authentication, scoped tools and tests.
- You bring
- An API or a database, and what an assistant should be able to do with it
- You get back
- An MCP server, authenticated, tested, documented and listed
- Shape
- Two milestones: read tools live, then write tools and the listing
Deliverables, method and fitaigrow.gg/aigrow-score(verifiable source, opens in a new tab)
RAG on PostgreSQL with pgvector, on your own documents
I build retrieval over your own documents on the PostgreSQL you already run, with pgvector, so answers come from your data and say where they came from.
- You bring
- Your documents, and the questions people ask of them
- You get back
- Retrieval in your Postgres, with sources and a measured hit rate
- Shape
- Three milestones: ingestion, retrieval with a test set, the application
LLM evaluation for a system that is already running
I tell you, with numbers, where the AI system you already run can be trusted and where it still needs a person, starting from what it records.
- You bring
- Read access to the system, its logs and its data model
- You get back
- A written report: what can be measured, what was, and what to change
- Shape
- Two milestones: the instrumentation review, then the measurement
Deliverables, method and fitOpenReview(verifiable source, opens in a new tab)
How the work runs
- Method
- A written brief in, a working deliverable out
- Cadence
- Milestones, each one approved before the next starts
- Meetings
- None required
- Clients
- United States, United Kingdom, Canada, German-speaking Europe
- Pricing
- Per milestone against a written scope, quoted per project
- Handover
- Code, deploy path, tests and the record of every decision
Why the work runs in writing
An asynchronous engagement is one where every decision, constraint and trade-off is written down at the moment it is made. It is slower to start and faster to finish, and at the end you own a readable record of why the system is the way it is.
Nobody has to be interviewed to find out why a decision was made, because the reasoning sits in the repository next to the code that implements it. Time zones stop mattering too, which is why most of the work is for teams in the United States, the United Kingdom, Canada and German-speaking Europe.
What this is not
No staff augmentation, no joining your standups, no hourly billing and no open-ended maintenance retainers. If what you need is an engineer inside your team every morning, you need to hire one, and I will say so before taking the engagement.
Questions about hiring
- Why are there no prices on this page?
- Because the figure depends on the scope, and a number without one would be a guess you could not hold me to. Send a brief and you get a written scope with a price for each milestone. If the project is not a fit for me, you get that answer instead, with the reason.
- What goes into a good brief?
- What the system has to do, what already exists, who will use it, and when it has to be live. A sample of the data helps more than a long description. Rough is fine: the scope I write back is where it gets precise.
- How long does a project take?
- A system that does one job properly is usually weeks, not months. The first working version of the 1o1.org dialogue engine existed in ten days; the product around it took a sequence of milestones after that. The variable is scope, which is why it is agreed in writing first.
- Can you work alongside our engineering team?
- For a defined piece with a clear boundary, yes: an agent, a connector, a retrieval layer, an evaluation. For work that means sitting inside the team day to day, no. In that case you need to hire, and I will tell you so.
- Who owns the code?
- You do, on delivery of each milestone, along with the deploy path and the written record of the decisions behind it. Nothing is held back as leverage and nothing depends on me to keep running.
Start with a brief.
Write what the system has to do, what exists already and when it has to be live. I read every brief myself and answer in writing, with a scope and a price for each milestone.
Or write to contact@guglielmo.io