Machine Learning Scientist
- Research & Development (Computational)
- Hyderabad, IN
Posted on March 19, 2023
Sure, text completion and image generation are cool, but do you know what’s cooler than training a transformer to power a ‘support’ chatbot with no actual power except the ability to frustrate online shoppers with smoother verbiage than the current generation of automatons, or fine-tuning Stable Diffusion to replace a stunt double’s face with Pedro Pascal’s when he can’t be bothered to show up on the set of The Mandalorian IV: The Merchandise Revenue Rises? Making the GPUs go brr in service of saving millions of lives by helping design breakthrough new biomedicines.
PopVax is an Indian biotechnology startup based in Hyderabad, and funded via research agreements with the Gates Foundation and Vitalik Buterin’s biosecurity company Balvi. For the past year, we have been building a novel second-generation mRNA platform for low-cost broadly-protective vaccines via computationally-driven antigen design, starting with an open-source booster vaccine candidate that is intended to protect against the entire sarbecovirus species and, if things go well, the betacoronavirus genus, including current and future strains of SARS-CoV-2, the virus which causes COVID-19.
At PopVax, we have seen up close that there are incredible capabilities in computational protein design just waiting to be unlocked that will enable us to pursue and perhaps even succeed at realizing wildly ambitious agendas such as preventing future pandemics via broadly-protective vaccines, tackling the deadly scourge of antimicrobial resistance, eliminating debilitating degenerative diseases like multiple sclerosis, and curing fiendishly tricky chronic conditions such as ME/CFS and Long COVID.
We seek to accelerate the development of these technologies while, at the same, ensuring that the benefits they bring are available to everyone that needs them as soon as possible. If all of the advances made in this seminal moment are controlled by western pharmaceutical companies, they will do what they always do – hold on tight to the IP while people in the developing world die. We must ensure that we in the rising world are at the vanguard of this computational revolution, lest we condemn our people to remain forever at the mercy of the same corporate overlords who denied Africa affordable HIV antivirals in the late 90s and continue to deny Indians lifesaving cystic fibrosis treatment – just two among uncountable examples.
We are building and refining a set of models using a variety of architectures – from transformers to graph neural networks to denoising diffusion models – to solve key problems in the design of vaccine antigens, targeted therapeutics, and nucleic acid delivery chemistry. We have already seen some impressive results from our work – early versions of our first vaccine candidate perform 10-120x better than existing mRNA vaccines in pseudovirus neutralization assays against key COVID-19 variants – but we need to expand our team with a small number of exceptionally competent and insatiably curious machine learning folks to keep the momentum going.
The ideal candidate need not have existing knowledge of biology, but needs to deeply understand current SOTA ML architectures, both in theory and implementation – we want someone just as comfortable reading the latest ICLR paper as implementing and training a large model on AWS. They must be willing to learn quickly, build fast, and act independently in an environment with a non-trivial amount of uncertainty, and be willing to rapidly change direction to wherever the data points.
If you come and work with us, we hope that in short order, its impact will accelerate work that saves or improves millions, and potentially billions, of lives. There aren’t many places that can offer you that kind of deep positive impact, not to mention the opportunity to work at the frontier of synthetic biology’s machine learning revolution.