One of the companies sponsoring the 2021st edition of Pi Day Coding Competition is InSilicoTrials, an Italian startup that allows pharma and medical devices companies to significantly decrease R&D costs (from 40% to 60%) and reduce regulatory approval times (from months to hours) thanks to its cloud-based platform for modeling and simulation. InSilicoTrials team is composed of “engineers and scientists with a passion for life sciences, digital innovation, and cybersecurity”, as explained by Alberto Palazzin, companies tech lead. “Every one of us contributes expertise and her experience to follow a shared goal: to de-risk and accelerate the research and development of medicines and medical devices by establishing the use of modeling and simulation in the healthcare sector. This is possible through cutting-edge technology made simple and cost-effective for the benefit of pharmaceutical companies, medical device producers, and people worldwide”.
What are the main technologies you use at InsilicoTrials?
We have a web platform on one hand, and scientific models offered on it on the other.
On the web side we have a so-called JAM stack with state-of-the-art technologies. The front end lives in the web browser and is built using the Javascript framework Vue.js. It is served via a content delivery network backed by cloud storage. The back end is a GraphQL API built on Node.js in a serverless fashion. Data is stored in storage as well as relational and NoSQL databases. Our infrastructure is hosted on the Microsoft Azure cloud.
The scientific models come in many different shapes and forms, from C++ to Python, R, MATLAB, or ANSYS. Every model presents its own set of challenges. For example, some require high-performance or parallel computing resources, while other models generate thousands of images or a large amount of data. To handle software provisioning with these many different technologies, Docker is a boon.
Why did you choose those technologies and what do you use them for?
The clear separation between the web client in the front end and an API in the back end helps enforce module boundaries and affords us a certain amount of flexibility on both ends. This way, we can build highly dynamic web interfaces with support for data visualization on top of an API that is easily tested and allows for composable workflows.
Tell us something about your developers and what are you looking for.
Due to the nature of working on a web platform that makes useable a heterogeneous mix of scientific models, InSilicoTrials is an environment that is welcoming both towards specialists as well as generalists. We think that just like a picture says more than a thousand words, a rapidly developed prototype enables more purposeful discussions than an overly detailed design document. At the same time, we are working in healthcare, and so we are aware of our responsibility — we strive for quality in our work and take verification and validation seriously.
What’s the most interesting thing you, as a developer, learned in working at InsilicoTrials?
Every modeling and simulation tool available on the InSilicoTrials web platform responds not only to product and software engineering requirements but also to the needs and considerations of our scientific partners, customers, as well as regulations. Working in an environment with multiple stakeholders like this was challenging at first, but when a product vision comes together, it can also be hugely rewarding.
