Dynamical characterization of multi-planet systems: I developed Bayesian analysis tools to infer the orbital properties of dynamically complex multi-planet systems. I use this information to look into these systems' possible formation histories and long-term orbital evolution.

• 55 Cancri: A five-planet system containing a transiting Super-Earth, pair of planets just outside of a 3:1 mean-motion resonance, a habitable zone planet, and Jupiter analog. Posterior samples available here.

• Gliese 876: A four-planet system hosting three planets in an orbital resonant chain. Allowing the planet mutual inclinations to float, we find the data favor coplanar configurations. Posterior samples available at here.

• ν Octanis: A mysterious binary system. The best known solution to the RV data has a Jupiter mass planet orbiting retrograde around the primary star.

Hot Jupiter formation: Do hot Jupiters form via disk migration or high-eccentricity migration? What if the answer is a little of both? And what if we can infer that from known hot Jupiter data? We apply the statistical modeling library Stan on these data to infer a mixture of populations. Full analysis available on Github.

RUN DMC: An efficient differential evolution Markov chain Monte Carlo to estimate exoplanet orbital parameters from RVs

MCMC importance sampling: A way to compute the fully marginalized likelihood by pairing an MCMC with an importance sampling algorithm. Code available at Github.

EPRV3 Evidence Challenge (in prep.): I ran a data challenge that asked teams to quantify the evidence for n-planets in RV data for set of simulated RVs and the same statistical model. More details and results are hosted at Github.