Download the current résumé (PDF)

Professional Summary

Physician-engineer working on production clinical AI and the systems that measure it. At Doctronic I design and ship capabilities inside a multi-agent medical harness (diagnostic interview, prescription refill, emergency surveillance) and build the evaluation infrastructure that gates them. Producing a plausible clinical agent is cheap; knowing whether it works is not, and that’s critical in medicine. Board-certified in internal medicine, pediatrics, and infectious diseases, with first-author work in computable phenotyping across 500M+ clinical observations.

Technical Skills

Agentic Systems: Multi-agent orchestrator/sub-agent architectures, Claude Agent SDK, deep-agent harnesses, tool-use design, deterministic safety gates and completion contracts, context engineering, latency optimization

Evaluation & Measurement: LLM-as-judge design and calibration, simulated-patient fidelity instruments, pre-registered study design, benchmark contamination detection, inter-rater reliability and paired significance testing; CRAFT-MD, HealthBench, MedQA-CS, OSCE/PACES rubrics

Retrieval: BM25, hybrid dense+sparse with RRF, HyDE, agentic retrieval loops, DSPy, GEPA optimization, gold-set construction, recall@k / precision@k

Programming & Analytics: Python, SQL, Polars, R, Docker, AWS (Bedrock, Fargate, S3), Azure AI Foundry, Google BigQuery, MCP servers, dsub, Google Batch

Statistical Methods: GWAS, PheWAS, phenotype risk scores, regression, propensity score matching, survival analysis

Healthcare Data: OMOP CDM, ICD, CPT, SNOMED, LOINC, RxNorm, ATC, FDA NDC, EHR phenotyping, cohort development

Experience

Senior Clinical Context Engineer

Doctronic | Mar 2026 – Present

  • Co-designing the next-generation consultation architecture: an orchestrator coordinating specialized sub-agents, patient thread decoupled from agent thread, and safety enforced through deterministic contracts rather than prompt instruction
  • Built the diagnostic interview capability (ranked differential with supporting and refuting evidence, plus next-best-question reasoning) and the gate that keeps any diagnostic statement from reaching a patient without a produced differential
  • Built the prescription refill flow mirroring the production state-regulated pipeline: renewal-safety interview, parallel safety engines over drug monographs, a monograph resolver behind a drift guard, and a deterministic fill-versus-escalate decision
  • Designed the guardian sentinel, splitting emergency surveillance out of diagnostic reasoning into an always-on parallel track
  • Found a contamination in a published benchmark where simulated patients disclosed their own artificiality to the diagnosing model, inflating scores; shipped a three-axis patient-fidelity judge as an upstream gate on every downstream result
  • Ran an end-to-end evaluation of clinical guidance retrieval, building a 300-case gold standard from real patient conversations; findings drove the architecture the production team adopted
  • Wrote or revised the clinical flag criteria for eight emergency presentations, with boolean flag-combination logic and per-question exemplars
  • Own technical interviewing and take-home design for clinical context engineering hires

Research Data Scientist

National Human Genome Research Institute, NIH | Aug 2023 – Mar 2026

  • Used Claude Code to orchestrate concurrent research projects across multiple disease domains (respiratory viruses, sarcoidosis, HPV-associated anal disease, long COVID, adverse drug events)
  • Architected distributed computing pipelines processing 155 survey-item PheWAS jobs and 26 ancestry-stratified GWAS using Google Batch/dsub
  • Developed phenotype risk score for long COVID prediction (AUC 0.94) that outperforms semi-supervised approaches, integrating wearable device data with EHR records
  • Built and validated computable phenotyping algorithms for 8 respiratory pathogens achieving 79-97% PPV; first-author publication in Scientific Reports (2025)
  • Led multi-investigator projects spanning EHR-survey phenotype comparison for genetic discovery, clinical risk factor identification, and genetic variant validation

Infectious Diseases Clinical Fellow

National Institute of Allergy and Infectious Diseases, NIH | Jul 2021 – Mar 2026

  • Managed complex infectious disease cases, building the clinical grounding that now informs my agent and criteria design work
  • Board certified in Internal Medicine, Pediatrics, and Adult Infectious Diseases

Medicine-Pediatrics Resident

University of Chicago Medicine | Jun 2017 – Jun 2021

  • Conducted retrospective EHR analysis identifying differential treatment responses in septic shock using machine learning-derived patient subgroups (American Thoracic Society Conference 2020)
  • Completed Summer Program in Outcomes Research Training (SPORT) fellowship

Graduate Student

NIH-Cambridge Scholars Programme, University of Cambridge | Aug 2010 – Nov 2016

  • Developed quantitative analysis pipelines for microscopy image data
  • First-author publication in Journal of Cell Science; co-authored Nature Cell Biology

Education

Degree Institution Year
Infectious Diseases Fellowship NIAID, NIH 2021–Mar 2026
Medicine-Pediatrics Residency University of Chicago 2017–2021
MD UT Southwestern Medical School 2008–2017
PhD, Clinical Biochemistry NIH, University of Cambridge 2010–2016
BA, Biology & Chemistry Texas Christian University 2004–2008

Selected Publications

Pradhan A, Waxse BJ, Matias WR, et al. Evaluation of four large language models on complex, infectious disease case scenarios. medRxiv. 2026. DOI

Fulda ES, Waxse BJ, Goleva SB, et al. 11 million days of longitudinal wearable data reveal novel future health insights. medRxiv. 2026. DOI

Waxse BJ, Bustos Carrillo FA, Tran TC, Mo H, Ricotta EE, Denny JC. Computable phenotypes to identify respiratory viral infections in the All of Us research program. Scientific Reports. 2025;15(1):18680. DOI

Waxse BJ, Rao S. Data Science for Pediatric Infectious Disease: Utilizing COVID-19 as a Model. Current Opinion in Infectious Diseases. 2025;38(5):493-498. DOI

Goleva SB, Williams A, Schlueter DJ, Keaton JM, Tran TC, Waxse BJ, et al. Racial and Ethnic Disparities in Antihypertensive Medication Prescribing Patterns. Clin Pharmacol Ther. 2024;116(6):1544-1553.

Full publication list on NCBI

Selected Presentations

  • Waxse BJ, et al. A Long COVID Phenotype Risk Score Reveals Decreased Steps and Elevated Resting Heart Rate Before and After Infection. AMIA Amplify 2026, Podium.
  • Waxse BJ, et al. Higher Step Count is Associated with Reduced Risk of Long COVID. IDWeek 2025, Podium.
  • Waxse BJ, Tran TC, Mo H, Denny JC. Identification and Validation of Common Respiratory Infections in All of Us. AMIA Annual Symposium, Podium. November 2024.

Contact

📧 bennettwaxse@gmail.com 📍 Denver, CO

GitHub LinkedIn