AI coding and Codex workshops for engineering teams
A hands-on workshop for software teams learning to use coding agents in everyday development. We work through how to give an agent useful context, define a task, inspect its changes, and verify the result.
Plan a team trainingWho this is for
Software engineers, technical leads, and product engineering teams. Participants should be comfortable reading code, using Git, and running their project’s checks.
When this is useful
- Your team has AI coding tools, but practices vary widely between engineers.
- Developers can generate code, but need better habits for context, testing, and review.
- You want a Codex workshop connected to actual engineering tasks and team conventions.
What we work through
Set up a useful task and context
Introduce a repository, its conventions, the expected behavior, and a bounded task. Decide what the agent should do and when a person should step in.
Work through the development cycle
Practice codebase exploration, implementation, debugging, and test development on a representative task.
Review and verify the result
Read the diff, challenge assumptions, run relevant checks, and recognize when an apparently complete answer needs more investigation.
Turn individual practice into team habits
Discuss shared instructions, review expectations, reusable workflows, and a small trial the team can evaluate after the workshop.
Working outputs
These outputs are a starting point for scoping. We agree the priorities together before the engagement.
- A worked engineering exercise with a reviewed diff and verification notes.
- Draft team conventions for task context, human review, and completion criteria.
- An initial practice plan for applying the workflow to the next development cycle.
Format and planning
A hands-on workshop or a series of applied sessions, in English or Turkish, remotely or onsite. We scope the duration around the exercises and team experience. Tool choice is agreed in advance; Codex can be the focus.
Before we start
Agree the team’s languages, tools, experience level, and approved AI access in advance. Use a sample repository or an approved, sanitized project. Accounts and any required licenses should be arranged before the session.
The experience behind the work
I am VP of Product, AI at Jotform, where I lead the AI Product division. Alongside product and engineering leadership, I contribute to company-wide AI transformation, hackweeks, practical training, and adoption programs. My advisory and training work draws on that operating experience.
My leadership experiencePractical questions
Does the training cover OpenAI Codex?
Yes. Codex can be the main tool for the workshop. The session covers engineering workflows such as understanding a repository, defining tasks, implementing changes, debugging, writing tests, and reviewing results.
Is this an official OpenAI certification?
No. This is an independently delivered workshop by Sabri Berkay Aydın, based on product and engineering leadership and AI enablement experience. It is not an OpenAI certification program.
Can we work on our own codebase?
Yes, when the repository and tools have been approved by your company and the exercise is appropriately scoped. A sample or sanitized repository is also suitable. Credentials and sensitive data should not be part of workshop materials.
Will using AI coding tools make our team faster?
The workshop does not promise a productivity percentage. The useful question is whether your team can complete suitable work more effectively while maintaining quality. A trial can compare cycle time, review effort, rework, and defects against a baseline.
Related work
Discuss the scope for your company
Share your company, the people involved, and the problem you want to address. We can work out the appropriate approach and scope together.
Plan a team training