AI applied to real products
I don't use AI as a decorative feature. I integrate it where it solves a concrete problem: reducing time, automating repetitive decisions, or scaling what a small team cannot do manually.
I identify which parts of your product benefit from AI and which are better solved with traditional code. Not every problem is an LLM problem.
AI quality depends on the context you give it. I design prompts, RAG, memory and data structures so the model returns useful output.
I integrate models, agents and data flows inside a clean architecture. Fullstack: from database to interface.
Every AI feature is instrumented. I track usage, cost, errors and satisfaction to iterate before scaling.
Systems that execute multi-step tasks: answer emails, generate content, coordinate workflows or act on data.
I connect your documents, data or APIs to language models for accurate answers based on your information.
OpenAI, Anthropic, Gemini or open-source models. I choose the right model per task, budget and latency.
Functional products in weeks where AI is the core differentiator from day one.
My own method to structure prompts, memory and context. Fewer hallucinations, more useful results.
Processes that used to require people now run with human supervision: classification, summarization, extraction and decision-making.
AI agent for freelancers that manages clients, projects, tasks, time and invoices through conversation.
View project →Platform that measures environmental and social impact using AI to process complex data and generate actionable reports.
View project →Nutritional assistant that recommends personalized recipes based on ingredients, allergies and user preferences.
View project →Smart notes for meetings: transcription, summary and action-item extraction with language models.
View project →Tell me what you want to solve. In a first conversation I'll tell you if AI adds real value or if another path makes more sense.
Schedule a conversation