Join our cutting-edge research initiative to architect autonomous software agents (actors) capable of self-replication and self-migration across untrusted and possibly unreliable infrastructure.

During this PhD internship, you will research techniques to design fault-tolerant, transparently persisting applications that can migrate across the infrastructure. The focus lies on continuing from (optionally verifiable) checkpoints and designing intelligent strategies to navigate unreliable, yet paid, host machines.
 

Duration: flexible, to be agreed (typically 3-4 months), starting time is flexible

Location: Antwerp (Belgium) 

Student enrolled in Ph.D. Computer Science/Engineering in Distributed Systems and/or Program Languages and/or Security Strong programming skills in JavaScript/TypeScript/Scheme or related easily serializable languages Language skills: English Experience in distributed systems, migration strategies, continuations, object capabilities, error-correcting codes, serverless computing, CRDTs and/or fault-tolerant systems are a big plus. A strong publication record is also a big plus.
  You will review scientific literature on object capabilities, capability machines, fault-tolerant computation such as the Ken protocol: https://web.eecs.umich.edu/~tpkelly/Ken/ You will develop various prototypes to improve the performance and strategies used for self-migrating agents. You evaluate techniques in the context of specific requirements.