Skip to content
§ services // what I offer

SERVICES.md

Project work for systems that have to hold up in production, and workflow automation for the manual jobs in between. Every service has a published starting price.

§ 01 // project work

Project work follows a defined process: a free first conversation, a paid scoping phase, then delivery in phases. see the method

build something new

from $500

A new production system from scratch: scoped, architected, built with tests and CI, documented and handed off.

I start from the problem, not a feature list. A paid scoping phase produces a signed spec with success criteria, an architecture sketch, milestones and what's out of scope. Delivery follows it in phases, with a written update every week and a demo at each milestone. The stack is chosen to fit the problem and to be run by your team after handoff.

what you get

  • a signed spec before any code is written
  • the codebase, with tests and CI/CD
  • a production deployment with error tracking and health checks
  • a runbook, API documentation and architecture decision records
  • full ownership of the repository

→ Teams with a defined problem and no system for it yet.

request this: build something new

fix what's broken

from $250

I take over a system that misbehaves, find the root cause, fix it and leave it observable and documented.

First I reproduce the problem and find out why it happens, not just where. Then I fix the cause, add tests so it stays fixed, and add the logging, error tracking and alerts that were missing, so the next problem shows up before your users report it. If the right fix turns out bigger than agreed, you get a written change order before I go further.

what you get

  • the root cause, explained in writing
  • the fix, with tests that cover it
  • logging, error tracking and alerts where they were missing
  • updated documentation and a runbook

→ Teams whose existing system is unreliable, slow or unmaintainable.

request this: fix what's broken

architecture audit

from $150

A review of your architecture, code and operations, delivered as written findings and prioritized recommendations. No implementation.

I read the code, the architecture and how the system is deployed and operated, and look for what will hurt first: reliability, security, data handling, cost and how hard the system is to change. Every finding comes with its impact and a recommendation, in order of priority, so you can act on it with or without me. Fixing what the audit finds is a separate engagement.

what you get

  • a written report of findings, each with its impact
  • recommendations in order of priority
  • a clear answer on whether to keep, scale or rebuild

→ Teams deciding whether to rebuild, scale or hand a system to someone new.

request this: architecture audit

AI / LLM integration

from $250

Adding LLM features to what you already run (assistants, extraction, reporting, MCP tools) with evals before deployment.

LLMs fail silently and drift over time, so I treat them as infrastructure, not magic. Each feature ships with an eval suite that measures it against real examples from your data, plus guardrails and fallbacks for when the model gets it wrong. Before building anything, I'll tell you if AI isn't the right tool for the problem.

what you get

  • the feature, integrated into the system you already run
  • an eval suite built from real cases
  • guardrails and fallbacks for wrong or missing answers
  • LLM observability: traces, cost and quality over time
  • documentation and a runbook

→ Products and internal tools that need AI to work reliably, not as a demo.

request this: AI / LLM integration

§ 02 // workflow automation

one-time automation

from $50

I automate a manual process once (scraping, data cleanup, document processing, migrations) and deliver the results with the script that produced them.

You describe the job and what the result should look like. I write the automation, run it, and check the output before handing it over. Typical jobs: scraping a site into a spreadsheet, cleaning and merging messy data, pulling fields out of a batch of documents, or moving data from one system to another. The price is fixed and agreed before I start.

what you get

  • the results, in the format you need (CSV, Excel or a database)
  • the script that produced them, so the job can be run again
  • short notes on how to run it

→ Anyone with a one-time job too tedious to do by hand.

fixed price: 50% to start, 50% on delivery

request this: one-time automation

automation pipeline

from $100

I automate a recurring manual workflow and deliver it running on your infrastructure, documented, with a runbook.

I map the workflow as your team does it today, then automate it with n8n, Python or both, depending on what your team can maintain. It runs on a schedule or a trigger, logs every run and alerts someone when a run fails, so a broken step never goes unnoticed. You can run it yourself or have me host it.

what you get

  • the pipeline, running in your environment
  • failure alerts and a log of every run
  • the source code, in your repository
  • documentation and a runbook

→ Teams repeating the same manual steps every day or week.

fixed price: 50% to start, 50% after its first successful run in your environment

request this: automation pipeline

hosting

Your pipeline runs on your infrastructure by default. If you'd rather not run it yourself, I can host it for you, from $15/month: I keep it running, monitor it and make small fixes when a source or API it depends on changes. New features are quoted separately.

Paid services the pipeline uses (LLM APIs, proxies) run on your own accounts and API keys, so their usage is billed to you directly.

tools

  • n8n
  • Python
  • Claude
  • Playwright
  • pandas
  • GitHub Actions

Not sure which one fits?

Send a brief anyway: choose "other" and describe the problem. I'll tell you which service fits, or whether none does.

send a brief