Fang-Ying Tsai

• H1B visa
• BNL Nuclear and Particle Physics Software Group (NPPS group member)
  - I developed a deep set neural network (DSNN) framework for VHbb (Higgs boson with a vector boson decaying into two b-jets) analysis. I was invited to give a 30-minute lecture followed by a 10 (eventually took ~30) minute Q&A at a summer school at Princeton University.
  - I am currently on Editorial Board on the llvv (2 leptons and neutrinos) analysis in ATLAS experiments.
  - I utilized DASK arrays to process large-scale data sets in a distributed computing environment.
  - I was able to test various distributed ML jobs across different schedulers, including HTCondor,
  SLURM, Ray cluster, GRID, and Google Cloud, to improve computation efficiency.
  Co-convener in the FastChain (Fast simulation in ATLAS) group (2 years till now)
  - I developed a TensorFlow ML model to improve the selection of high pT events for reconstruction in the track-overlay simulation workflow, achieving an estimated 45% speed-up. Leading developer and author.
  - I am a technical supervisor on the Track Overlay project.
  - I am driving innovation in fast simulation in ATLAS group.

• I developed a novel data-driven method to estimate the ZZ → llvv background, which is the dominant
  background in two analyses papers. (The method is limited by data statistic. It would become
  very helpful in HL-LHC.)
• I was invited to be a convener for the MET group in 2022 in ATLAS. I was told that I'm the great asset of the MET group by two convenors multiple times.
  - I developed and optimized algorithms to prevent double counting of electrons and jets in the reconstruction of MET calculation, resulting in a 30% improvement in the accuracy of MET calculation.
  - Investigated increased number of events with high MET in Z → ee events and tracked it down to a problem in the e/𝛄 reconstruction code and then came up with a short-term prescription to mitigate the problem.

- Research (CMS experiment) performance and skills learned from work helped me to get a position in Germany. During the interview, I was told that I was already qualified at the PhD level, and I received the offer the next day.

VPON is a Taiwanese company that provides an advertising platform for app publishers to monetize their applications. Their platform integrates ad services with technical support for both Android and iOS, utilizing JavaScript as a bridge to facilitate communication between mobile apps and backend servers. I worked as an engineer to provide a technical support between clients and developers.