{{title}}
{{description}}
Homepage anchor for site.
As ML workloads grow in complexity and scale, domain-specific languages (DSLs) and specialized hardware have become essential tools in research and industry—enabling domain experts to express intent more directly and achieve significantly higher performance. At the heart of these innovations lays the compiler infrastructure. Modern compiler technology bridges the gap between high-level DSLs and low-level, performance-critical execution on heterogeneous hardware. With the rapid rise of new DSLs and diverse hardware targets, the demand for efficient, maintainable and robust compiler tooling has never been greater. My current research focuses on building the next generation of compiler infrastructure—systems that are not only performant, but also maintainable and robust, enabling rapid prototyping of new DSLs and hardware targets. By leveraging principles from the functional programming paradigm, we are building an MLIR-compatible compiler framework that streamlines DSL development and promotes reuse of common compiler concepts, creating a shared library of abstractions and tools.