Service · AI
AI integration for iOS apps
I add AI to iPhone apps and to the processes around them: language models through the Claude API, RAG pipelines on your own data, Core ML models on device, and automation that connects LLMs to business systems. I use the same tools every day in my own development.
- +40%
- user engagement with AI data analysis
- 10+
- clients with RAG automation
- −60%
- UI prototyping time with AI agents
What I build
- In-app features powered by LLMs: assistants, summaries and insights from the user's own data.
- RAG pipelines that let a model answer from your documents and databases.
- Core ML models that run on the device, for speed and privacy.
- Analysis of data coming from connected devices, from raw readings to a clear next step.
- Automation with Make.com, Zapier and Python cloud functions that connects LLMs to business data.
How I approach it
- Pick the right place for the modelOn device with Core ML when data is sensitive or latency matters, in the cloud through the Claude API when the task needs a large model.
- Ground it in your dataRetrieval over your own content, so answers are specific and can be checked.
- Protect the dataEncryption and access rules designed in from the start. I built full in-app encryption of financial data at Atom to support SOC 2.
- Measure itShip behind a flag, watch usage and quality, then extend what works.
Where I've done it
At Technogym I introduced AI data analysis and AR training features that increased user engagement by 40%.
As a freelancer I've built automation that connects LLMs to business data through RAG pipelines for more than 10 clients, and a Python AI microservices architecture of over 20 cloud functions with 99.9% uptime.
I also use agentic AI every day: a development environment built on Claude CLI generates SwiftUI components and cut UI prototyping time by 60%. At Doona AI-assisted workflows cut the development cycle by 45%.
Technology
Questions
Can you add AI to an app that's already on the App Store?
Yes. I usually start with one feature that uses data the app already has, release it to a group of users, and measure the result before going further.
Does the user's data leave the device?
It depends on the feature. Core ML runs entirely on the device. When a cloud model is needed I send only what the task requires, encrypted, and document it for your privacy policy.
Do you work on AI outside the app too?
Yes. I build automation and RAG pipelines for business processes, and I can set up AI-assisted development workflows for your team.
Have a project like this?
Tell me about it. I'll reply with a first technical take and a timeline estimate.