About Me

Hello! I’m Chiril Calin, and I am currently an AI/ML Engineer at Pfizer. I work primarily with our Vaccines R&D teams to tackle interesting pLLM problems like epitope prediction, and our Vaccines Operations teams to solve optimization problems like minimizing assay re-runs.

My primary interests include:

• Machine Learning: We can create models to neatly understand and explain our world.
• Interpretability: We should create models that can be easily understood and explained.

To the point of machine learning, my work is generally centered around the creation and fine-tuning of protein LLMs, and other architectures which allow scientists to predict and model otherwise intractable relationships within biology/chemistry. I hope to share this work primarily in the form of publications, and will aim to open-source as much of my work as possible.

My interest in interpretability is much broader, and I work to understand model thinking both within biology and in general. If we can understand the way our models think and structure concepts, we should be able to meaningfully steer our models towards desired outputs and away from harmful ones. I aim to share this work primarily in the form of accessible blog posts, so that people will be able to follow along and quickly iterate on the work.

You can find my code on GitHub. If you have any questions or comments, please reach out to me at chirilcalin7@gmail.com.