Julia performance tuning for live application

We are looking to contract with a Julia performance specialist who can help improve the performance of our julia based machining-AI application that powers Toolpath.com.

We use julia for the main body of the app, though we do have some modest C and C++ libraries we compile against as well. We do have one major dependency on pyOCC (python wrapper of OpenCascade) which we call through a python interface.

We leverage a lot of computational geometry and spatial searching algorithms. Its all tied into a AWS lambda based application. We have some amount of brute-force lambda parallelization capabilities as well.

We have a full test suite, with over 90% coverage.

We’ve built out a set of basic performance profiling tools, including native julia based tools along with sentry.io based logging.

All of these would be at your disposal, along with the ability to work with our primary development team to help pinpoint and fix bottlenecks.

We’re looking for gains across our whole application stack. We expect they will come from a combination of algorithmic improvements, code changes to improve memory usage, increased parallelization. We are also interested in potential GPU based speed improvements, but consider these to be our lowest priority.

Along with the general test suite, we can provide a set of specific test cases that are proving to be computationally challenging.

We’re willing to pay $80 an hour for the consulting work, along with bonuses for achieving greater than 10x improvements on any of the critical test cases we provide, or on the overall cost of our test suite.

Please reach out to justin.gray@toolpath.com if interested.


We’ve filled this position!