VAICRO
Co-founded VAICRO to replace expensive, time-consuming clinician-administered assessments with real-time AI voice interviews. CTO and engineering lead. Looking for engineers who want to build with us.
enth.net
Building software that accelerates brain health research.
Bay Area · CTO & Co-Founder, VAICRO · Director of Engineering, Brain Health Registry
I started my career as a staff research associate — running MRIs and processing brain images. When I saw how much time we were losing to manual data management, I started building tools. Each project grew out of a real problem the lab was facing, and that pattern has defined my career.
Over 25 years at UCSF and the VA, I've built systems that support research at scale: a participant recruitment platform serving 100,000+ registrants, neuroimaging pipelines that have tracked hundreds of thousands of cases, and an AI voice interview platform that conducts research assessments by phone.
I studied Molecular & Cellular Biology with an emphasis in Neurobiology at UC Berkeley. That science background still shapes how I build — I think in terms of research workflows, data integrity, and what scientists actually need.
Co-founded VAICRO to replace expensive, time-consuming clinician-administered assessments with real-time AI voice interviews. CTO and engineering lead. Looking for engineers who want to build with us.
Director of Engineering for the Brain Health Registry and the VA Advanced Research Center. Lead a small team building research platforms and infrastructure used by investigators across the country.
Research platform with 100,000+ registered participants. Self-service registration, consent, questionnaires, cognitive testing, and remote biospecimen collection. Has generated 250,000+ participant referrals to 100+ studies. Adopted by ADNI and others.
Lets investigators deploy and administer AI voice interviews with custom prompts and post-interview analysis. Manages real-time voice conversations between participant and AI agent, records audio and transcripts, and runs investigator-defined post-processing.
Co-developed an AI-administered version of the CDR with Washington University collaborators — translating the traditional clinician- administered instrument into structured AI interview modules with domain prompts, follow-up logic, and automated scoring.
Python framework that lets scientists define neuroimaging processing pipelines executable by research staff, with full tracking of inputs, outputs, outcomes, and automated next steps. Used to build 10+ pipelines tracking 200,000+ cases across 50+ projects.
High-performance parallel processing of DICOM images: duplicate detection, coding verification, PII stripping, hierarchical storage, and database indexing. Has processed 250,000+ scans comprising 30+ million individual DICOM files.
Manage the VA Advanced Research Center's imaging lab infrastructure: VMware virtualization, Dell storage, Linux workstations and processing nodes, FreeIPA for centralized identity, and an SGE-based processing grid.
The throughline across everything below: lean on the platform, keep the surface small, let composition do the work.