Memoro

Memoro – AI Conversation Intelligence

Memoro is an AI-driven platform that transforms spoken conversations into structured knowledge, automatically generating transcripts, summaries, and action items. Visit the live product: memoro.ai

From Pitch to Prototype

My journey with Memoro began at a startup event in Constance (Germany), where I met founder Till Schneider. Impressed by his pitch, I joined as the Founding Engineer for the critical initial phase. My mission was to shoulder the technical responsibility during the first months, transforming a visionary concept into a functioning reality.

Engineering the AI Pipeline (MVP)

While the concept was clear, the technical path was not. My core task was to architect and build the backend orchestration pipeline that powers the AI analysis. I developed a robust system focused on reliability and fault tolerance:

• Self-Healing Architecture: I engineered the logic to handle the instability of external AI services. If a process failed—for instance, due to exhausted API tokens—the system automatically paused and resumed once resources became available.

• State Persistence & Checkpoints: I implemented a system to save “intermediate artifacts” at every step. This ensured that expensive computations (like transcription) were preserved even if the subsequent analysis step failed, preventing data loss and reducing costs.

• API Orchestration: I connected various disparate APIs into a seamless workflow, ensuring raw audio data was correctly processed into structured JSON output.