# Arkadij's Talk Rehearsal

## Introduction slide

- ETH: you can do a self-deprecating joke about how ETH is the Swiss MIT but <something funny></something>
- Super smooth, keep it that way.
- Emphasize that all of the things you're talking about are public domain at this point, so there's nothing proprietary

## Enzyme engineering

- Put a title slide before the "concept & innovation points" slide, just to visually demarcate the transition to the first project.
- Thus far, really smooth, great stuff.
- IRED test case overview: increase the size of the figure, if possible. Also, put source for the figure.
- Prime audience by saying "hierarchical Bayesian estimation model"
- "multiple measurements" -> you can consider the term "replicate measurements".
- Candidates to prioritize (slide 8) -- missing the "probability above a threshold value".
- Traditional enzyme eng: don't forget to highlight that there were some intricacies in decision-making that led to such a rapid enantioselectivity shift.

## Antibody Engineering

- V. good explanation of screen vs. counter-screen
- Goals of thesis: don't skimp on this! :smile_cat:
- The way you're communicating your thesis motivation is really good: "clustering" sets expectations to be something scikit-learn, and then bam! you give them a probabilistic model :smiley_cat:
- Comparison to other models: "I'm familiar with how these models work, but in the interest of time won't go into them, happy to talk about that later."
- You probably can re-structure the intro to the "how we cluster" slide by summarizing it as, "something, something, scikit-learn.fit_transform()"
- Your explanation of "mixture of multinomials" is really good.
- "...has been applied to a wide range of projects and targets."
- MMM applied to panning outputs of H-CDR length 13: you probably want to communicate the idea that we're using these sequences to validate that the model is doing the right thing.

## Moderna Job Talk

- Strongly emphasize Antibody engineering talk
- Then 5-10 minutes about jax-unirep, highlighting software engineering skills in context of first novartis project

- Bold, curious, collaborative and relentless