These are the situations that most often lead a company to bring us in, from a room full of curious engineers to a portfolio of legacy apps waiting to be modernized.
Many organizations have a few developers experimenting with AI and many more who haven't started. A single full-day session gets everyone working from the same playbook at the same time. Your team sees the same techniques applied to the same problem, which makes it far easier to build shared habits afterward. It also gives managers a common language for talking about AI-assisted work.
Customer Reviews
I was astonished that, in such a short period of time, something that other companies had taken months to understand, this company was able to execute. Since then, we have never looked anywhere else. I would never consider another company. This is how much we believe in them.
The other agencies we looked at, frankly, wanted to tell us how great they were. Sembit instead asked a lot of probing questions and provided the kind of consultation that really made us comfortable in hiring them.
It's been excellent every step of the way. The initial conversations were excellent — we had an instant human connection. In addition to being professional, they also had a human touch that's irreplicable.
I've really loved working with them. Sembit always delivers what they say they're going to deliver, in the time they say they'll deliver it. Whenever we have a deadline for a task delivery, the team gives us detailed and frequent updates about their progress. It's been a great partnership.
Sembit's team really took on the mindset of our business which resulted in a better product with thoughtful decisions from developers. We have been able to have healthy discussions about the pros and cons of a design choices which ultimately makes a better product in the long run.
Sembit has over-delivered on their promises at every step of the process. They've come in ahead of schedule and below budget, and they let us know if we are asking for too much. They are clear communicators, and it made a world of difference for our company.
Sembit's Claude Code training teaches engineering teams by working through real projects on their own code.
Teaching an Engineering Team Claude Code on Real Code
An established lumber company brought us in to train its engineering team, and about 30 people filled the room. Shaun Davidson, Sembit's VP, taught by working through a real project live: using Claude Code to modernize SalesForecast, one of the company's 13-year-old legacy apps, while sharing practical techniques and answering the team's questions. The day continued with a strategy session on what's coming in AI-assisted development and a discussion with senior technical leadership.
Sembit trained a 30-person engineering team on Claude Code by modernizing one of their own legacy apps live, then led a strategy session on the future of AI-assisted development.
Why Sembit?
Projects completed
Years in business
Average Years Experience per Employee
Sembit is an Oregon-based software team with deep, hands-on experience building real software with AI. Our engineers use Claude Code to design, build, and modernize applications, so we teach from what we've learned shipping production work rather than from slides or vendor demos. Beyond coding agents, we bring broad AI expertise across strategy, business automation, and AI platform integrations, which helps us answer the hard questions about where AI fits and where it doesn't. Because we are a smaller, senior-heavy team, your training is led by experienced engineers, including our founders, who understand both the technology and the business decisions behind it.
Let's plan a training session built around your team.
Talk with Nathan Theophanes (Sembit's President) and Shaun Davidson (Sembit's VP) about your team, your goals, and the kind of project we could tackle together live. We'll be in touch in hours, not days.
The core of the training is real, hands-on work on a project chosen with your team, such as modernizing a legacy application or building a new custom app. Your team watches each step, from initial investigation through planning, execution, and verification. Working on familiar code and real business problems makes the techniques concrete. It also shows honestly what the agent does well and where an engineer needs to step in.
Claude Code works best when it understands your project. We show how to give it the right context through project instruction files, conventions, and documentation, and how to structure a repository so the agent can navigate it. Good setup up front saves hours of correction later.
Large tasks like porting an app need a plan before any code is written. We demonstrate how to break work into steps, review the plan, and then let Claude Code run for extended stretches while keeping it on track. This is the difference between a quick code suggestion and a morning's worth of real progress.
AI-generated code still needs to be checked. We cover how to review changes efficiently, how to have the agent write and run tests, and how to compare old and new behavior side by side. Verification is also where bugs and security issues in legacy code come to light.
Enterprise teams need control over what an AI agent can access and do. We discuss permission settings, working safely with databases and cloud resources, and how to fit Claude Code into your existing review and deployment process. The goal is speed without giving up oversight.
AI-assisted development is changing quickly. Our strategy session covers what is practical today, what we expect next, and how it may affect hiring, team structure, delivery timelines, and your technology roadmap. We leave plenty of time for your team's questions.
FAQs About Our Claude Code Enterprise Training
A full day in two parts. The morning is a hands-on tactics session where we tackle a real project live with Claude Code and share practical tips along the way. The afternoon is a strategy session covering where AI-assisted development is heading, followed by open Q&A. Many clients also add a separate discussion with senior technical leadership.
While this is what a typical training looks like, we will shape the format, length, and project around your team's goals.
Yes, with the right process and review. Claude Code is very capable, but an experienced engineer still needs to guide it, review its work, and verify the result. We teach exactly that.
Yes. Our training is designed to be delivered in person, with your team in the room, so people can watch the work closely and ask questions as it happens. Contact us to discuss the format that works best for your organization.
Yes. Many teams want help applying what they learned, whether that is building new applications, modernizing existing ones, setting up team standards for AI-assisted development, or planning a broader AI strategy. We offer ongoing support across our development and AI strategy services, and you can reach our founders directly.
We recommend your own code. We work with you ahead of time to choose a project that fits your goals, such as modernizing a legacy application, building a new custom app, or tackling a feature your team has been meaning to get to. Working with your real code, data, and infrastructure makes the training directly relevant, and your team leaves with real progress on real work. If policy prevents using a production codebase, we can use an internal or sample project instead.
Both. The hands-on work is aimed at developers, while the strategy discussion is valuable for engineering managers, technical leadership, and executives. We are also happy to meet with senior management separately to talk through planning, staffing, and risk.
We work within your policies. Before the session, we agree on which codebase, data, and cloud resources the agent can access. During training, we cover permission settings and guardrails so your team can use Claude Code safely in its own environment afterward.
No prior experience is required. The session works well for full engineering teams with a mix of skill levels, and managers and executives are welcome to sit in. Developers who have never used an AI coding tool and those who already use one daily both come away with something practical.
