João Rocha

JTML Interpreter | Creator & Maintainer | [GitHub](https://github.com/yourusername/jtml)

_C++-based reactive templating engine with WebSocket synchronization for dynamic web apps_
- Designed a thread-safe dependency graph with Tarjan's cycle detection for reactive variables.
- Reduced UI update latency by 40% via debounced batch recalculations.
- Implemented TLS-secured WebSocket server handling 500+ concurrent connections.

- Tech: C++17, WebSocket++, CMake, Docker, GitHub Actions.

• Developed and maintained ML pipelines using FastAPI and Argo CI workflows on a Kubernetes cluster
• Developed and maintained the Forecasting project data ingestion and preprocessing pipelines, and customer facing API for model serving.
• Maximized efficiency of metadata processing using PostgreSQL to handle over 1TB of data
• Improved ML workflows through custom DAG pipelines using Pandas, enhancing efficiency and data processing
• Developed a Streamlit based internal application to streamline customer onboarding, reducing time from days to hours

• Designed and deployed ML models for predicting electricity generation in wind and solar farms using ARIMA, h2o, TensorFlow, and Prophet
• Improved data exploration and cleaning using R and Python
• Augmented operational efficiency by building real-time prediction pipelines for renewable energy
• Achieved 40% reduction in deployment times and 30% increase in system reliability by leading backend infrastructure setup using Docker and custom bash scripts for CRON jobs
• Gained 50% boost in data throughput and 20% decrease in operational costs using Ray Cluster