Research & Projects
My research spans programming languages, runtime systems, concurrency, and AI-based software development. I focus on language and runtime designs that make programs safer and easier to write, and on practical tools and systems that put them to use.
Current Research & Development
Lightscale AI
Current - Founded 2026
Platform that generates business software from natural-language descriptions. Large language models write programs in a domain-specific language designed for them, which expresses only the domain logic; database, service layer, authentication, and user interface are generated. Certain error classes are excluded by language design, and the compiler statically checks further selected architectural properties, such as access control and data integrity rules.
Internet Computer & Motoko Language
2022-2025 - DFINITY Foundation
Core contributor to the Internet Computer blockchain platform and Motoko programming language. Designed and implemented runtime features including incremental garbage collection and enhanced orthogonal persistence for smart contracts.
Incremental Garbage Collection for Blockchain
2022-2023 - DFINITY Foundation
Incremental garbage collection for Motoko smart contracts on the Internet Computer, designed for deterministic execution and bounded instruction limits per message. Published at VMIL 2023.
Enhanced Orthogonal Persistence
2023-2024 - DFINITY Foundation
Orthogonal persistence for Motoko smart contracts that retains the program state in main memory across contract upgrades, avoiding explicit serialization to stable memory. Published at VMIL 2024.
Major Research Projects
Parallel Checker for C#
2015-2021 - HSR Rapperswil / OST
Static analysis tool detecting data races, deadlocks, and other concurrency issues in C# directly in the IDE. Features high precision and fast analysis times. Open-sourced under MIT license.
Composita Programming Language
2004-2014 - ETH Zürich & HSR Rapperswil
Component-oriented programming language featuring hierarchical composition, dynamic wiring, fine-grained concurrency, and communication-based interactions. Supported by a dedicated operating system kernel for large numbers of concurrent components with fast context switches.
GPU Parallelization in .NET
2014-2016 - HSR Rapperswil
Dataflow programming model and runtime system for .NET that maps high-level dataflow graphs to GPU code, in collaboration with QuantAlea. Aims to spare developers low-level GPU-specific programming.
Cloud Task Parallelization
2012-2016 - HSR Rapperswil
.NET runtime extension that ships ordinary shared-memory tasks, together with their reachable code and data, to high-performance computing clusters in the cloud, and merges their side effects back into local memory.
Persistent Oberon
2003-2007 - ETH Zürich
Extension of the Oberon language with integrated orthogonal persistence, avoiding explicit persistence mapping. Program state is kept in an automatically managed object store and survives system restarts.
Compiler & Runtime Systems Book
2018-2021 - HSR Rapperswil / OST
Textbook on compiler and runtime system design. Hands-on approach building both compiler and virtual machine from scratch, covering interpreters, JIT compilation, object-oriented runtime support, and garbage collection.
Industry Collaboration
In addition to academic research, I have conducted applied research projects with industry partners in areas such as concurrent programming, static analysis, and runtime systems.