Interview: Elliot MacGowan, COO at Agnostiq

As one of the leading sponsors at Quantum.Tech Boston 2022, we caught up with Elliot macgowan, COO at Agnostiq who explored Agnostiq as a business and what they do, what motivated Agnostiq to build Covalent and what they are looking forward to for Quantum.Tech Boston.

For those who are not aware of Agnostiq; are you able to give us a bit of an insight into your business?

Agnostiq develops software tools and applications that aim to make quantum and high performance computing resources more accessible to enterprises and developers. Agnostiq is the team behind Covalent, an open source workflow orchestration platform designed to help users manage and execute tasks across heterogeneous compute resources (CPUs, GPUs, QPUs and more). You can learn more about Covalent on our website and on our Github. We are based in Toronto, Canada with team members all across North America.

What can you tell us about Covalent?

Covalent is an open source Pythonic workflow tool used to execute tasks on advanced computing hardware. It is designed to make it easy for users to keep track of their computationally heavy experiments by providing a simple and intuitive framework to store, modify, and re-analyze computational experiments.

Among its many features, Covalent enables users to:

• Rapidly iterate and prototype exploratory research models
• Automate, manage, and share reproducible experiments
• Visualize data and task dependencies with an interactive graphical user interface
• Exploit knowledge of task dependencies to effortlessly parallelize tasks across multiple cores and distributed computing resources
• Run code in heterogeneous compute environments, including in hybrid-cloud and hybrid-quantum configurations
• Understand where time and money is spent across a project

Covalent is rapidly expanding to include support for a variety of cloud interfaces, including HPC infrastructure tools developed by major cloud providers as well as emerging quantum APIs.

How does it have the potential to optimize customer workflows?

Customer workflows are becoming more heterogeneous in nature. They utilize a growing list of hardware types (CPUs, GPUs, QPUs), software languages (Julia, C, Python) and frameworks (CUDA, PennyLane, Qiskit) to support them. Workloads are also increasingly moving to the cloud, meaning customers need to be aware of all the services available to them to help manage and benefit from the transition. It also means that many jobs have become hybrid, where some part of the customer’s workflow is done locally or using on-premise clusters.

While the net result of all this innovation is extremely positive for customers, there is still a lot of work that needs to be done to make sure all of the new tools and modalities are being used optimally. Covalent is designed to help reign in this complexity, and to provide a simplified interface for the world’s most heterogenous workflows. With features like checkpointing, scheduling, monitoring, logging and alerting, Covalent brings the ideas of DevOps to the scientific community, but with an easy-to-use, lightweight interface, requiring only minimal changes to existing code.

What potential real world applications do you see for Covalent?

Covalent will be invaluable in any computationally intensive setting when using classical or quantum hardware. Wherever HPC is used, Covalent will empower scientists, engineers, business modelers, finance professionals and others to submit, disaggregate, monitor and repeat their experiments and numerically intensive compute jobs more easily and effectively. Any researcher running large scale simulation or optimization workflows, or training machine learning models will benefit from Covalent.

What motivated you to build Covalent?

Covalent started out as an internal project, built to help get our internal research team up and running in a cost-effective manner. In the beginning, we faced the dual challenge of wanting our research team to use the best hardware and software available, without getting bogged down optimizing infrastructure. We surveyed the landscape and found that there were no tools available that could give us everything we needed in one package so we decided to roll our own. It did not take us long to realize that all other teams would soon face the same challenge. We saw more teams building their own DIY tools, which of course took their time away from their core, differentiated work of developing new algorithms and/or applications. Our hope is that Covalent will free them up from this undifferentiated work, and allow them to focus on what they are truly unique at.

You are kindly one of our sponsors at Quantum.Tech in Boston this June; what are you looking forward to at the conference?

Agnostiq has exhibited at Quantum.Tech conferences in the past. We are always impressed at the excellent planning and execution that goes on behind the scenes to make such a successful conference. We are excited to soak up the ambiance, meet great people across the spectrum from technical to business-minded, attend some great talks on cutting-edge material, and just have fun in the beautiful city of Boston. Please be sure to stop by our booth and say hi - if nothing else, we have incredibly soft t-shirts for the taking.