AI Lab is a place for projects currently in motion: applications, automation, model experiments, image, video, data, and smaller technical trials that may later become separate products or services.

The point is not one technology or one narrow use of LLMs. Language models can act as an interface, a production tool, an automation layer, a prototyping partner, or one element inside a larger system. Each project tests a different part of that space.

Heart Hackers is a dating application developed as a full product: from backend architecture and mobile app flows to messaging, push notifications, user profiles, visual direction, publishing, and promotion.

The current direction is the iPhone version and further refinement of the user experience.

Related page: Full-stack

Video experiments focus on automating short artistic and promotional forms. AI is used not only to generate images, but also to search for rhythm, mood, narrative movement, and coherence between text, motion, sound, and editing.

Related page: Video

Szukacz is a job-offer agent-scraper and a tool for making sense of the job market. Technically, the core is written in Python; Playwright is used for selected job portals, LLMs evaluate the match between a CV and a particular offer, and Hermes, through MCP and WhatsApp, facilitates daily use. It is built around collecting, filtering, classifying, and preparing data so that real opportunities, repeated patterns, and blind alleys can be recognized faster.

Technology: Python, Playwright, OpenAI API, MCP, Hermes

Literary Models are experiments with fine-tuning language models on Polish literature from the public domain. The aim is not to create a universal chatbot, but to study style, tone, rhythm, and the possibility of building models rooted in a specific body of texts.

Related page: Fine-tuning

Festinga is a return to a service dedicated to festivals of the world. The site collects places, events, rituals, music, traditions, and seasonal calendars of culture.

AI can help rebuild the content, organize data, support translations, and give the project a clearer structure for further development.

Website: Festinga

AI Lab is not a catalogue of finished products, but a place where ideas are tested by time, attention, and return. The abundance of possibilities has become its own challenge: suddenly almost everything can be built, but only a few things deserve to be stayed with longer. Projects made only for tomorrow usually fade the day after; the ones that survive are given focus, patience, and care for meaning. AI can accelerate the beginning, but it does not replace the decisions that give things permanence.