Hi! You found R²!

Hey, I’m Russell.

I build production-grade ML and agentic systems for NLP, retrieval, and forecasting in healthcare, pharma, e-commerce, aerospace, and beyond.

I’m a PhD student at UC Berkeley in the Department of Bioengineering, advised by Dr. Reza Abbasi-Asl and funded by the NSF GRFP.

Current work:

  • Evidence-guided clinical trial design: Measuring whether trials adopt published recommendations for broadening patient eligibility across 500K+ trials. 19K+ assessments across 658 recommendations and 4,400+ trials.
  • AI platform @ FenixCommerce: Built and delivered 4 production systems for retail clients, coordinated by an autonomous agent platform. +25% on-time delivery (+60% relative) on 5.96M orders. $4.93M incremental revenue across 7 retailers.
  • Agentic workflow manager: Multi-agent orchestration framework that powers both industry work and PhD research. Coordinates parallel workstreams, manages task dependencies, and automates deployment.

Past work:

  • Adversarial robustness of medical LLMs: Testing how medical AI models fail under challenging clinical scenarios. (arXiv)
  • High-throughput screening system for cardiovascular drug phenotyping @ Novartis. (article)
  • End-to-end test engineering for internet satellite systems @ Amazon.
  • Thermal testing system for hemostatic diagnostic devices @ Werfen. Discovered overheating flaw affecting blood sample integrity.
  • MitoGAN: Synthetic image generation for mitochondrial dynamics research.
  • Computational experiments for identifying CAR-T protein candidates.
  • ScanAlert: Electronic medication monitor for improving medical adherence.

Honors:

Full CV