Rafael Ravedutti

- Worked with database kernel development, focusing on code optimization. - Experience configuring and managing instances on Huawei Cloud.

PhD student and research assistant at the Chair of Computer Science 10 - System Simulation - Developed MD-Bench, a proxy-app used to evaluate the in-core performance of MD kernels from several state-of-the-art packages - Created and developed P4IRS, an intermediate representation and compiler for particle simulations that efficiently generates particle simulation solvers (such as MD and DEM) for multi-CPU and multi-GPU targets - Performance analysis - Code optimization for multi-CPU and multi-GPU clusters

- Implemented distributed-memory parallelism with MPI for MD simulations - Performance and scalability analysis - Software development with AnyDSL (using Impala language, a dialect of Rust) - Code optimization and portability - Comparison between AnyDSL code and traditional implementations in general purpose languages (miniMD) - Design and implementation of a domain-specific library using the AnyDSL framework that uses higher-order functions to express interaction kernels for particle simulations

- Software development in Python with the Django framework, using PostgreSQL database. - Refinement, documentation and optimization of existing application. - Agile methodology with GitLab for control versioning and task organization.

- Big data applications using both no-relational (MongoDB) and relational (PostgreSQL) databases - Python (Django framework), Taiga, JIRA, Slack, Scrum - Experience configuring and managing instances on IBM Cloud - Micro-services architecture and configurations using Docker containers - Significant performance improvement for applications that need to load massive amounts of data - Generic solution to work with data from different sources and evaluate medical protocols using compiler construction techniques

- Administration and management of Linux, Solaris and BSD compute servers. - Experience configuring and managing instances on AWS Cloud. - Network configurations and maintenance (routing, security, virtual networks). - Installation and configuration of Solaris application and storage zones with Apache and OracleDB. - Automation scripts with Bash. - Development of infrastructure panel using PHP and Oracle database.

- Developed and maintained free software applications in different projects - Programming in Python with Django, Ruby on Rails and PHP (Elgg), using PostgreSQL database. - Automation scripts with Bash and version control with Git

Chair of Computer Science 10 - System Simulation - Thesis submission and defense still pending - Developed complex code-generation solutions to efficiently solve physical simulations on multi-CPU and multi-GPU targets - Worked with performance analysis and modeling of Molecular Dynamics kernels - Efficiently implemented simple and clean state-of-the-art performance strategies from MD simulation packages efficiently (through reverse-engineering their codes) using SIMD intrinsics (AVX2, AVX512)

I did part of my masters in the University of Erlangen-Nurnberg. This included attending lectures and working as a research assistant in the Chair of Computer Science 10 - Systemsimulation - Classes: Advanced programming with C++, Computer graphics, GPU programming, code optimization, distributed-memory parallelization with MPI. - Developed a domain-specific library for fast and scalable molecular dynamics simulations on multi-CPU and multi-GPU clusters.

The dissertation was developed towards an extension in the Halide domain-specific language to improve its profiling capabilities. - Studied compiler construction, optimization and parallel programming techniques. - Performance analysis with PAPI and Likwid. - Reverse engineering to understand undocumented parts of the Halide DSL. - Developed an extension in the Halide DSL to improve its profiling capabilities.

- Studied DSL approaches to generate optimal code for heterogeneous hardware (CPU/GPU). - Developed an image processing domain-specific library using the AnyDSL framework and compared its performance against Intel OpenCV. -Provide a comparison study between DSL solution and general-purpose language implementation (OpenCV), wrote my monograph on the subject.