ZdamLEK – AI supercharged learning for medical students
Background
The medical profession demands not only rigorous academic knowledge but also the ability to apply it accurately under high-pressure conditions. For young doctors in Central Europe, the Lekarski Egzamin Końcowy (LEK) is a pivotal, high-stakes licensing examination that determines the course of their career. ZdamLEK is a robust AI-powered eLearning platform helping hundreds of students and young doctors.
Traditional Approach
The traditional method of preparation relies heavily on manually sifting through extensive medical textbooks, lecture notes, and official question databases provided by the central examination board (CEM).
Information Overload: The sheer volume of medical knowledge required to pass the exam is immense, leading to burnout and inefficient study cycles.
Verification Challenge: Students spend critical study time manually cross-referencing information from practice questions against dozens of official sources, a process that is slow and prone to human error.
Lack of Reliability: While digital resources are available, many lack the critical feature of source verification, leaving students to question the factual basis of explanations—a dangerous proposition in a medical context.
Existing e-learning platforms offer easier to digest form of lectures and presentation, but lack the individual approach and tools enhancing the learning process.
Technological and Factual Challenge
The primary technical hurdle was the integration of a sophisticated AI agent that must operate with absolute fidelity to medical facts, effectively overcoming the notorious “hallucination” issue prevalent in general Large Language Models (LLMs). This required XTRN to engineer a highly specialized, literature-grounded AI system capable of:
Retrieving and synthesizing complex medical information exclusively from a pre-vetted, official knowledge base.
Generating comprehensive answers, schematics, and comparative tables.
Crucially, providing source citations (bibliographies and links to official literature) for every piece of information provided, thus adhering to the highest standards of factual knowledge.
User Expectations Challenge
Beyond accuracy, the platform needed to deliver a superior, modern user experience. Potential customers, tech-savvy medical students, expect:
Seamless accessibility across devices (desktop and mobile).
Intuitive dashboards for real-time progress tracking.
Advanced quizzing and mock exam functions that allow for highly granular, targeted learning (e.g., focusing only on previously difficult or unreviewed questions).
XTRN’s challenge was to synthesize a massive, fact-critical medical knowledge base with cutting-edge AI and a world-class user interface, all while ensuring the highest caliber of software development.
Benefits
High-Quality AI Assistant
XTRN developed an AI agent that is explicitly not a general LLM. This specialized agent retrieves and explains medical concepts, procedures, and differential diagnoses with guaranteed factual accuracy, as every answer is instantly backed by direct citations and links to official medical literature. This adherence to regulatory and factual standards proves XTRN’s ability to develop mission-critical custom solutions.
Time Savings and Efficiency
Students no longer waste hours searching textbooks. The AI assistant and over 12,000 high-quality, authored explanations provide instant, reliable clarity, allowing students to dedicate their time solely to learning and retention.
Comprehensive, High-Fidelity Content
The platform is built around the most current CEM 2025 database for the LEK exam, ensuring students are studying from the official, relevant question set, supplemented by XTRN’s meticulously created reviews and comments.
Personalized Learning
The built-in dashboard and learning planner provide powerful analytics, enabling students to track their progress, identify knowledge gaps, and strategically plan the revision of the entire question base, thereby optimizing their study path and improving confidence.
