So you like computational science, and you want to leave academia?

Roughly once a year, my undergraduate thesis advisor (Prof. L. Sriramkumar) directs students with a scientific background and an interest in computational science to reach out to me. This is usually because the student is interested in exploring the world outside of academia, and they usually think their software skills are the only thing useful in the "industry". It's disheartening that academia doesn't equip students pursuing basic sciences to understand career possibilities in the industry, but I digress. I've had such conversations 10s of times since 2016, so it's about time I wrote this down.

Let's start with your skills

What is Computational Science?

Please note that this advice is based on my personal experience and is specific to students with an interest and experience in Computational Science. Computational Science, or Scientific Computing, applies computational techniques to solve scientific problems. For instance, I used the Python programming language and packages from the Scientific Python ecosystem to analyse astronomical data. Similarly, I also used the Mathematica programming language to simulate models of the universe. Don't ask me about it because this happened a long, long time ago in a galaxy far, far away.

Depending on the college you pursued your undergraduate studies in, you might have received formal training in Computational Science. For instance, I remember pursuing courses on "Numerical Methods in Programming" and "Introduction to Computational Physics". "Foundation of Computational Physics" is currently a first-semester course at my alma mater, and courses on "Computational Physics" and "Computational Science in Engineering" are available online at NPTEL.

One of the failings of Computational Science courses in colleges is that they don't inform students about software best practices. For instance, the assignments I submitted prioritized accuracy and precision of scientific results that the code generated, and didn't check whether or not the code was organized into modular and reusable functions. Reading code I wrote during my undergrad makes me wince - it's all one giant Python file with spaghetti code. If you want to improve your software and computing skills, I would strongly recommend going through the Software Carpentry and Data Carpentry material, and the MIT "Missing Semester of Your CS Education" course.

Where can you go?

Research Software, Data Science, and everything in between

A few Academic Institutes and Universities have recognized the importance of research software. Berkeley Institute of Data Science (BIDS) at UC Berkeley and the eScience Institute at the University of Washington were absolute pioneers, paying people to write research software full-time within academia. US Research Software Engineer Association (US-RSE), Society of Research Software Engineering (RSE-UK), The Carpentries, and the NumFOCUS job boards share full-time computational science opportunities in the western hemisphere regularly.

Unfortunately, the ecosystem in India isn't as mature at the moment. Given my background in Astronomy and Astrophysics, I know that every once in a while, the Inter-University Center for Astronomy and Astrophysics (IUCAA), the Raman Research Institute (RRI), and the Laser Interferometer Gravitational-Wave Observatory (LIGO)-India have scientific computing opportunities (Research Assistant). Occasionally, Free/Libre and Open Source Software for Education (FOSSEE) at IIT Bombay hires folks to work on scientific computing FOSS projects.

Moving further away from academia, there are Scientific Software Developer roles in the industry, like the one at Enthought that I lucked into. There are only a handful of companies that do this kind of work, such as Quansight Labs in the US and Quantstack in the EU. More Scientific Software Developer roles are in-house, for instance, KLA-Tencor India, Tokyo Electron (TEL), and KBI Biopharma. Startups like Skyroot Aerospace, Agnikul Cosmos, GalaxEye Space, and more are hiring folks who are able to marry scientific computing with domain expertise. Check out other portfolio companies of Venture Capital firms like Speciale Invest to get an idea of the variety of domains where scientific computing work is being applied.

Farthest from academia are Data Analyst and Data Science roles, in product and services companies. More often than not, these involve writing scripts to clean and prepare datasets, train, test, and deploy Machine Learning models, monitor deployed models for drift, and so on. Shoutout to Tiger Analytics, MathCo, and Fractal Analytics because I know folks at these firms. These roles will be the hardest to get into because of the sheer magnitude of competition that you will be facing, especially with the handicap of a "basic sciences background".

A welcoming community?

The SciPy US and EuroSciPy conferences are premier conferences for Scientific Computing, especially in the Python ecosystem. If you're in the West, you should seriously attend your nearest SciPy conference. Not only will you run into people of your own kind, but you will also get to network with people who might hire you. SciPy US and EuroSciPy sponsors are usually a who's who of Scientific Computing organizations, so you should go through as many old conference websites as you can. If Data Analytics and Data Science are your jam, you could consider participating in PyData communities and conferences around the world.

SciPy India was the premier conference on Python for Education and Scientific Computing in India. Unfortunately, the conference has been on hiatus since 2021, after being on an impressive run since 2009. A few of us are reviving the SciPy India community and making it a welcoming space for conversations around scientific computing across programming languages and domains. We have organized a few virtual community calls where people from various backgrounds talked about Computational Science in their fields. We also organized a miniconference on FOSS in Science on the sidelines of the IndiaFOSS 2025 conference. Say Hi to us online on Zulip or consider speaking at one of our upcoming virtual community calls.