Work with me
Senior engineering for the part that cannot stay vague.
I work directly with founders and technical leaders on web systems where architecture, execution, and control matter. You get the engineer doing the thinking and the implementation—not a sales layer.
Good fit
Bring me work that needs judgment.
- a web product with unclear or fragile architecture;
- a focused system or subsystem that needs an accountable builder;
- an inherited codebase that has become hard to change;
- a team introducing agents without durable context or verification;
- a prototype that must become a credible engineering foundation.
Probably not a fit
I am not a commodity delivery layer.
- bulk page production with no engineering problem;
- anonymous staff augmentation with no decision authority;
- promises of fully autonomous software development;
- work selected only by the lowest hourly price.
01Clarify before committing
Architecture & agent-readiness review
For an existing product, repository, or planned system where the next technical move is uncertain. I examine the product intent, architecture, constraints, development context, and realistic role for agents.
You leave with
- a concise map of the system and its load-bearing risks;
- the most valuable intervention to make next;
- agent-readiness gaps in context, package knowledge, authority, and verification;
- a scoped recommendation for implementation.
Design and deliver
Focused build or rescue
For a web application, integration, internal tool, prototype, or difficult subsystem that needs senior hands-on ownership. I can work across browser code, Node.js services, data, third-party APIs, and Linux deployment.
The working principle
We define the valuable outcome and boundaries first. I then implement in inspectable increments, using agents where they improve speed while keeping decisions, tests, and responsibility explicit.
Increase engineering capacity
Agent-driven engineering system
For a product or team that wants more than ad hoc chats with coding agents. I establish the durable context and control surfaces that let agents contribute coherently over time.
Possible scope
- product-to-code cognitive context using ADSM;
- repository instructions and authority boundaries;
- agent-friendly package and architecture documentation;
- verification, tests, and review checkpoints;
- bounded event-driven workflows, including GitHub orchestration where useful;
- handoff that leaves the system understandable to your team.
How the work runs
Direct, explicit, and evidence-led.
- FrameWe identify the costly problem, desired outcome, constraints, and authority.
- ModelI make the product and architecture decisions explicit enough to execute.
- BuildI implement and use agents for bounded work where they create leverage.
- VerifyTests, observable behavior, and review replace confident guesswork.
- TransferYou receive working software and durable context, not dependency on one chat.
A useful first message
Describe the system, the painful problem, and the outcome you need.
Add the current stack, constraints, and what has already been tried. I will respond with the most sensible way to begin.