Visito AI
- Architected and scaled the core platform infrastructure, leading a full cloud migration from Azure to Google Cloud Platform, enabling a more reliable and cost-efficient foundation for production workloads and CI/CD automation.
- Designed and implemented a production-grade AI orchestration layer, integrating OpenAI (tool calls) for conversational workflows and Google Gemini for high-volume document processing, supporting context-aware, multi-step AI interactions at scale.
- Built Retrieval-Augmented Generation (RAG) pipelines using MongoDB Vector Search and embeddings, significantly improving response relevance and reducing hallucinations in AI-driven customer interactions.
- Re-architected the real-time communication layer from a tightly coupled system to an event-driven architecture using RabbitMQ and WebSockets, eliminating latency bottlenecks and improving reliability under concurrent user traffic.
- Developed scalable data ingestion services in Python, combining asynchronous workers, message queues, and embedding pipelines to process and enrich large volumes of unstructured data.
- Owned cross-platform frontend architecture, enabling a unified web and mobile experience through shared component libraries and consistent engineering standards across Next.js and Ionic-based applications.
- Acted as technical owner from MVP to scale, translating business requirements into system design decisions while balancing feature velocity, technical debt, and long-term platform stability in a remote-first startup environment.