Skip to main content
Work with me

Work with me

AI integration

I lead AI integration projects as the product manager, with my team building. I start by checking where AI earns its place and where plain rules work better, then we build the feature so a person signs off on risky actions and the product has a fallback when the model fails.

Tell me about your projectSee what we built

01 / Audience

Who it’s for

  • Products that want an LLM feature they can rely on
  • Teams with manual review or triage work that eats their week
  • Data and ML teams that need quality checks before training

02 / Deliverables

What you get

  • Use-case assessment: where AI earns its place in your product and where rules are the better tool
  • Prototype and evaluation: a working prototype checked against your real examples
  • Production integration: the feature inside your product, with a log of what the model did
  • Deterministic fallbacks: the feature keeps working when the model is slow, unavailable or switched off
  • Model choice: hosted models, local models or bring your own key, picked for your data and constraints

03 / Process

How we work

  1. Find the use case

    We look at the work your team or users do today and find the step where AI saves real effort without adding risk.

  2. Prototype and evaluate

    We build a prototype and run it against your own examples, so we judge it on your own data.

  3. Integrate with guardrails

    The feature goes into your product with a confirmation step wherever an action is risky.

  4. Launch and adjust

    After launch we watch where the model helps and where it fails, and adjust prompts, rules or scope.

04 / Built with my team

What we built

05 / Questions

Common questions

Which models do you work with?

I pick the model for the job. In ADDY, my open-source personal AI assistant on GitHub, rules handle the requests they can and a hosted LLM (Groq, falling back to NVIDIA) handles the rest. Local models or a key you bring are options when your data needs to stay with you.

Does my data leave our systems?

That depends on the model. If your data has to stay with you, we can look at a local model running in your own environment during the use-case assessment.

Can you add AI to a product we already have?

Yes. We work inside your existing codebase and add the feature alongside what is already there, so the rest of the product keeps working as it does today.

What if AI is not the right fit?

Then I will say so. Some problems are solved better with rules or a cleaner workflow, and the assessment is there to find that out before anything is built.

How much does an AI integration cost?

It depends on the use case and how deep the integration goes. Share your budget in the contact form and I will suggest a scope that fits it.

Work together

Have a project in mind?

Tell me what you want to build, and share the scope and budget you have in mind.

Tell me about your project