Skip to content
Constantine Dzik

Practical AI · Claude Code & Codex · Local models

Learn practical AI by building real things.

Hands-on learning in coding agents such as Claude Code and Codex, AI agents, practical LLM workflows, running local models, and AI-assisted software development.

Designed for engineers, technical professionals, and anyone who wants to move beyond basic prompting and learn how AI tools can be applied to real work.

Explore learning options Contact me Get updates →

Learn by doing.

The Academy is a practical learning space focused on using modern AI tools in real workflows.

Instead of only explaining concepts, sessions are built around examples, guided exercises, coding, experimentation, and complete workflows.

The goal is simple: understand how the tools work, where they are useful, where they fail, and how to apply them responsibly in everyday technical and professional work.

Practical

Work with realistic examples and real workflows.

Hands-on

Learn by building, testing, and experimenting.

Flexible

Choose group sessions, workshops, or private lessons.

Tool-aware, not tool-dependent

Learn Claude Code, Codex, and local models while focusing on principles that remain useful as tools change.

Learning options

Choose the format that matches your goals and experience level.

For engineers and technical professionals

Coding Agents for Real Projects

Learn how to use coding agents such as Claude Code and Codex as part of a real software-development workflow—from understanding an unfamiliar repository to planning, implementation, testing, debugging, code review, and automation.

Topics

  • Repository exploration
  • Architecture understanding
  • Project instructions and agent configuration
  • Planning before implementation
  • Feature development
  • Refactoring
  • Debugging
  • Test generation
  • Git and code review
  • MCP integrations
  • Skills, hooks, and subagents
  • Comparing Claude Code and Codex on the same task
  • Running local models for coding
  • Safe and controlled AI-assisted development

Example outcome. Use a coding agent to analyze a project, implement a feature, create or update tests, review the changes, and build a reusable AI-assisted development workflow.

Ask about coding-agent lessons

Group learning

AI & Agentic Workflows Group

Small-group sessions focused on practical AI, LLM workflows, agents, tools, retrieval, automation, running local models, and the engineering decisions behind useful AI systems.

Topics

  • How modern LLM systems work
  • AI agents and tool use
  • Retrieval and context
  • Prompt and context design
  • Workflow automation
  • Model evaluation
  • Hosted models versus local models
  • Running open-weight models locally
  • Reliability and failure modes
  • Human review
  • Privacy and security
  • Practical implementation patterns

Example outcome. Participants learn how to design and evaluate an AI workflow and build a small practical project or automation together.

Ask about group sessions

1:1 learning

Private AI Lessons

Individual sessions tailored to your experience, goals, and current project.

Suitable for

  • Engineers learning Claude Code or Codex
  • Technical leads exploring AI adoption
  • Founders validating an AI idea
  • Professionals who want structured guidance instead of a generic course

Possible topics

Claude and Claude CodeCodexAI-assisted codingLLM fundamentalsAgentic workflowsRetrieval and RAGEmbeddings and vector searchAI evaluationMachine learning fundamentalsComputer visionRunning local modelsLocal and private AIAI project architecture

Private lessons can be structured as a single session or as a short learning program built around a specific goal or project.

Discuss a private lesson

Who this is for

Engineers

Use AI tools to understand, build, debug, test, and maintain software.

Technical leads

Understand how to evaluate and introduce AI-assisted development and AI workflows into engineering teams.

Founders

Understand how to turn an AI idea into a realistic architecture and prototype.

Curious professionals

Build a practical understanding of AI without getting lost in hype or unnecessary theory.

Topics we can explore

Possible areas for lessons, workshops, and group sessions—not a fixed catalogue.

ClaudeClaude CodeCodexAI coding agentsAI-assisted software developmentAI agentsAgentic workflowsLLM systemsRetrieval and RAGEmbeddingsVector searchModel evaluationPrompt and context engineeringComputer visionAnomaly detectionRunning local modelsOpen-weight modelsPrivate and on-premises AILocal-first AIAI architecture

Example: coding-agent session

From repository to working feature

A practical guided session showing how a coding agent—Claude Code, Codex, or a local model—can be used throughout a software-development task.

  1. 01Understand the repository
  2. 02Identify the relevant architecture
  3. 03Create or refine project instructions
  4. 04Define the task
  5. 05Build an implementation plan
  6. 06Make controlled changes
  7. 07Run tests
  8. 08Review the diff
  9. 09Identify risks and improvements
  10. 10Document the final workflow

The exact content can be adjusted to the participant’s technical level and project.

Ask about this session

Example: AI Agentic Workflows Group

A small-group session focused on understanding how AI agents work and how to design useful, bounded agentic workflows.

Topics may include

LLM + toolsAgent planningTool callingWorkflow stateRetrievalHuman approvalEvaluationFailure recoverySecurity boundariesCost and latency

Outcome. Build and review a small agentic workflow together.

Private lessons built around your goal

Not everyone needs a full course.

Private lessons can focus on one specific problem, tool, or project—for example:

Sessions can be technical or non-technical depending on the learner.

Contact me about private lessons

How learning works

Live and interactive

Sessions focus on active discussion, demonstrations, and hands-on work.

Project-based

Where possible, learning is connected to a real or sample project.

Flexible depth

Topics can be explained at beginner, intermediate, or advanced level.

Public-safe examples

Use public, synthetic, or approved materials. Do not use confidential company data.

Reusable outcomes

Participants leave with notes, examples, templates, code, checklists, or workflows they can use later.

About the instructor

Constantine Dzik is a technology lead with experience across software engineering, data systems, machine learning, and applied research.

His work includes practical AI systems, LLM-based workflows, embeddings, vector search, AI-assisted engineering, and running local and open-weight models.

He has experience evaluating AI development tools, supporting AI adoption across engineering teams, and helping professionals use Claude Code, Codex, and related AI tools in practical workflows.

His dissertation, “Modeling and Analysis of the Operational Characteristics of Solar Panels in Photovoltaic Power Plants,” focuses on anomaly detection using operational telemetry, physics-informed digital twins, statistical methods, and autoencoder neural networks.

Learn more about my background →

Interested in learning together?

If you are interested in Claude Code or Codex, AI agents, running local models, a small group session, or private lessons, send me a short note about what you want to learn or build.

Contact me Get Academy updates

Your details are sent securely to my inbox. If your browser blocks it, the button falls back to a pre-filled email draft instead. No third-party tracking, and no sensitive information is requested.

Frequently asked questions

Do I need programming experience?

No. Some topics, such as Claude Code and Codex, require basic software-development knowledge, while private and group sessions can also be designed for non-technical learners.

Can I book a private lesson?

Yes. Private lessons can be tailored to a specific tool, topic, or project.

Do you offer group sessions?

Yes. Small-group sessions can focus on coding agents, practical AI, LLM systems, agents, local models, or related topics.

Is the Academy only about Claude?

No. Claude Code and Codex are important practical tools, but the broader focus is practical AI, LLM systems, agentic workflows, running local and open-weight models, computer vision, and practical implementation.

Can we use my company project during a session?

Only if you are authorized to do so. Public, synthetic, or approved materials are preferred. Confidential employer or customer information should not be shared.

Is this an official Anthropic or OpenAI program?

No. This is an independent educational initiative and is not affiliated with or endorsed by Anthropic, OpenAI, or any other technology provider.

Are sessions recorded?

This depends on the session and will be agreed in advance.

Practical AI Systems Academy is an independent educational initiative. Product and company names are trademarks of their respective owners. The Academy is not affiliated with or endorsed by Anthropic, OpenAI, or other technology providers mentioned in its materials.