Timothe Petitjean

Leading and developing the AI backend for a real-time voice assistant for Alzheimer's patients and their caregivers. Designing a conversational architecture in Python integrating LiveKit (low-latency bidirectional audio), multi-model LLM orchestration with automatic fallback, and multi-provider STT/TTS pipelines. Developing conversational intelligence components: context management, critical keyword detection (distress), dynamic personalization, and automatic structured summary generation. Implementing event-driven services: autonomous call scheduling, intelligent triggering, real-time alerts via Firebase Cloud Messaging. Deployment and 24/7 operation on OVH Cloud: high availability, monitoring, error management, and system resilience. Participating in product iterations and technical presentations to business angels.

Designing an experimental protocol for the study of associative memory in Drosophila melanogaster. Developing a vision tracking system with feature extraction and statistical analysis on Raspberry Pi. Supervising 2 interns.