Hire me
Hire an AI integration engineer
Add AI to the product you already have: summaries, search, assistants and automation built into your backend, data and user interface, without a rewrite.
Quick answer
As an AI integration engineer I add AI features to existing apps and SaaS products: I connect the OpenAI and Anthropic (Claude) APIs to your backend, data and user interface, with RAG for answers from your content and tool calling into your own APIs. I build this today on the AI platform Joni (joni.ai), on top of 5+ years of production Node.js backends.
What I build
LLM features in your app
Summaries, smart search, writing help, classification and data extraction where your users already work.
AI assistants and copilots
An assistant inside your product that answers from your data and takes actions through your APIs.
RAG over your content
Embeddings and vector search, so answers come from your documents and records, with permissions respected.
Tool calling into your APIs
Validated calls from the model into your own endpoints, with approval for anything risky.
Model choice, cost and speed
The right model per feature, caching, streaming and usage limits so costs stay predictable.
Testing and monitoring
Test sets before release and tracking of quality, latency and cost in production.
Results on real products
Joni (joni.ai)
AI product platform
My main ongoing work: backend services, agent workflows and AI integrations for a live AI product.
Resume Builder โ CV Maker AI
AI in a shipped app
An App Store app with an AI assistant that improves resume text and exports ATS-friendly PDFs.
CV Maker โ Resume Builder
200K โ 1M+ users
Scaled a live app backend with about 60% faster APIs, the kind of backend an AI feature has to fit into.
Tools I use
OpenAI APIAnthropic APIRAGTool callingVector searchLangChainLangGraphNode.jsTypeScriptPythonMongoDBRedisAWS
How a project works
- Scope call. We talk through the goal, the current state of the product and what success looks like.
- Written plan. You get fixed milestones, delivery dates and a clear price before any work starts.
- Build in small pieces. I ship in small steps with regular progress updates, so you always see what is built.
- Tested delivery. Delivery includes tests, documentation, handover and free support afterwards if something breaks.
Frequently asked questions
What does an AI integration engineer do?
Connects AI models to an existing product: the backend calls, data access, prompts, user interface and the testing and monitoring around them, so the AI feature works for real users.
Can you add AI without rebuilding our app?
Yes. AI features plug into your current backend and frontend through new endpoints and components, so the rest of the product stays as it is.
Which models do you integrate?
The OpenAI and Anthropic (Claude) APIs, chosen per feature on quality, speed and cost, with retries, timeouts, streaming and rate-limit handling.
Is our data safe when we use AI APIs?
I send only the data a feature needs, keep keys on the server, respect user permissions in retrieval, and log usage. Check the provider data terms for your contract as well.
Can the AI answer from our own data?
Yes, through RAG: your content is indexed with embeddings and vector search, and the model answers from what is retrieved instead of guessing.
Can the AI take actions in our product?
Yes, through tool calling into your APIs, with strict schemas, validation and human approval for anything destructive.
How do you control AI costs?
Caching, smaller models where quality allows, trimmed context, batching background work and per-feature cost tracking.
How long does an AI integration take?
A focused feature such as summaries or smart search often takes days to a few weeks. An assistant with RAG and actions takes longer. You get a plan with milestones first.
Which stacks do you work with?
Node.js, TypeScript and Python backends, with MongoDB, Redis and AWS. I follow your code style, reviews and release process.
Can we start with one small AI feature?
Yes. Starting with one feature that is easy to measure is usually the best way to prove value before a larger build.