Not at #HWO25 but had to watch the Youtube stream to see the coolest version of the HR 8799 orbit movie from @bmacastro.bsky.social

We jump on the differentiable modeling bandwagon in order to fit millions of degrees of freedom in our optical model. This allows us to measure drifts in optical alignment that happen when JWST slews between targets. Because of this, we can do better than unsupervised data-driven methods like KLIP.

Excited to share our interdisciplinary work on developing new methods to subtract off the glare of the star to reveal fainter exoplanets. We show in simulation that we can build a full optical model of JWST/NIRCam and generate images of the star to subtract off and reveal faint planets (1/3)

I actually followed through on a new year's resolution!

I may have been thinking too much about optics recently because the first thing I thought of when I saw these truck lights is GMT pupil.

Realized I never shared this when I was at OHP: this is the now-retired ELODIE spectrograph that discovered 51 Peg b. You can see the fibers that inject both starlight and calibration lamp light, the echelle grating that dispersed the light, and the detector that captured the spectra.

Let's see how the algorithm here likes coronagraphs. Can you see the small black dot at the center of this optic? (you might need to zoom in) That is used to occult the light of a star to see planets around it.

I had the opportunity to make a pilgrimage to the Observatoire de Haute-Provence and see both the telescope and the spectrograph that discovered 51 Peg b. Super cool and awesome that the telescope is still detecting new exoplanets!