The Applied AI Research Fellowship at Evozyne is designed for researchers who want to stress-test ambitious ideas against one of the most challenging frontiers in applied AI today: generative design under real-world biological constraints. As a Fellow, you will work on foundational questions in representation learning, generative modeling, and optimization, with the opportunity to see your ideas evaluated against real experimental outcomes and translated into therapeutic programs.
Your work will directly influence how Evozyne evaluates, evolves, and deploys its generative AI models for protein design.
Who You AreYou're excited by problems where the data is messy, the constraints are real, and the path forward isn't obvious. You thrive in ambiguity, and you're motivated by applying your work to real-world scientific challenges to see how your ideas hold up in practice. You are already operating at the leading edge of applied AI and want to push your thinking further by applying it to complex, high-impact challenges in drug discovery.
What You'll Be InvestigatingAs an Applied AI Research Fellow, you will help drive the evolution of Evozyne's generative AI design platform. Example research areas include:
- Benchmarking generative protein models, including Evozyne's own, on their ability to produce functionally diverse and biologically meaningful designs.
- Evaluating the value of integrating large-scale metagenomic resources (e.g., Global Ocean Gene Catalog) into current internal database.
- Exploring alternatives and extensions to Bayesian Optimization such as Knowledge Gradient, Entropy Search, and related methods for multi-objective optimization problems.
- Developing and applying deep learning approaches for remote homology detection
- Investigating multi-family VAE models to enable protein design when sequence support is limited or when optimizing phenotypes across protein families.
These efforts are intended to surface failure modes, challenge assumptions, and directly inform how Evozyne designs proteins and advances therapies.
Education + Experience- Late-stage PhD student or postdoc in a quantitative or computational field
- Hands-on experience applying AI/ML to complex, real-world or scientific datasets
- Experience working on problems where data is noisy, incomplete, or difficult to interpret
- Familiarity with modern machine learning approaches (e.g., deep learning, generative models, or related methods)
- Evidence of meaningful contribution to research, open-source work, or applied projects
- Exposure to interdisciplinary work (e.g., biology, chemistry, physics, or other scientific domains) is a plus
Few places offer the combination of proprietary experimental data, real therapeutic programs, and the freedom to explore foundational AI questions under real biological constraints. If you want your best ideas tested where they matter most, and the chance to help redefine how AI is applied to protein design, we'd like to connect.
