BACKEND ENGINEER, DISTRIBUTED SYSTEMS & SIMULATION INFRASTRUCTURE
Backend engineer focused on distributed and real-time systems. Experience building APIs, networking services, online infrastructure, simulation tooling, observability systems, and data pipelines operating under real-world load. Strong in reliability, performance, and end-to-end system design.
Designed and implemented runtime-configurable AI training infrastructure for simulation-based autonomous agent development. Built with C#, Python, JSON, and YAML files. Managed environment configuration, episode and curriculum stepping, agent observations, and reward flows.
Implemented a deterministic record/replay system for agent observability using a format based on comma-separated-values (CSV). Enabled precise replays with the addition of tools like frame-stepping and time-scale adjusting to simplify debugging.
Developed and deployed backend infrastructure including matchmaking servers in C#, REST APIs with TypeScript and the Node.js runtime, and server stress tests using C#.
Designed NoSQL documents and served data for a live-service software product from MongoDB using the Mongoose object-document-mapper (ODM).
Co-authored and maintained a proprietary content management system (CMS) for organizing and scheduling the delivery of live-service content to clients across Google Cloud's content delivery network (CDN).
Developed real-time spatial visualization systems for Apple Vision Pro, ingesting MLB StatCast data and presenting live baseball events in 3D space.
Scripted and optimized spatial user interfaces (UI) for navigating baseball content and live game data.
Authored scripts for contextual asset loading based on user viewing preferences.
An end-to-end, bidirectional, real-time, climate telemetry system. Data is collected by a custom-built, embedded system for an ESP32 microcontroller and streamed to a web dashboard. WebSocket connections managed by a Node.js server deployed to Cloud Run on Google Cloud.
A demo of AI-powered semantic validation. Supported by a TypeScript on Node.js backend and MongoDB database, this interactive experience leverages LLMs to validate open-ended responses against user meaning rather than looking for exact matches.
For opportunities or inquiries, reach out via LinkedIn or email me at christopher.ravosa99@gmail.com.