Dual-Energy X-Ray Absorptiometry (DEXA): Principles, Clinical Applications, and Relevance to Minimally Invasive Surgery
First author. Forthcoming in Laparoscopic, Endoscopic and Robotic Surgery, 2nd ed., 2027
CV
Experience
- Software Team Lead, Ardurra - Houston, TX - Jan 2026 to present
- Lead software delivery for internal teams and municipal clients, including the City of Houston’s water utilities; hands-on in architecture and implementation.
- Embedded with the City of Houston’s water utilities: shipped 40 operational dashboards in five months, replacing vendor Power BI with self-hosted HTML/JS apps backed by Python, DuckDB, and Parquet pipelines on S3.
- Reverse-engineered a 104-page, 12.5-million-row Power BI compliance model (5.5 GB of exports) into a DuckDB/Parquet pipeline (196 MB) powering a 58-page static web app, and validated the migration plant-by-plant against 12,000+ regulatory violation counts.
- Consolidated three vendors’ sensor BI platforms (about 72 tables across three semantic models) into a single canonical data model with automated, lock-protected incremental publishing.
- Founder & Lead Engineer, Basil Health - 2025 to present
- Built and operate a clinical reporting platform that turns DEXA body-composition scans into personalized patient reports: in production at three paying clinics, 300 patients processed.
- Sole engineer across the stack: reverse-engineered the scanner’s Access database, wrote the Postgres ingestion pipeline, and ship guardrailed LLM-generated reports; re-platforming onto FHIR/Medplum with byte-parity tests between the Python and TypeScript insight engines.
- Shipped a full-stack patient portal verified by 58 API tests and 19 Playwright end-to-end specs.
- Researcher, The Robotic Global Surgical Society (TROGSS) - Sep 2026 to present
- AI engineering lead on a study of video models that detect robotic suturing errors across surgical platforms (in development).
- Software Engineer, Ardurra - Houston, TX - Mar 2024 to Dec 2025
- Led a two-engineer team delivering about 10 access-controlled RAG deployments (Azure OpenAI, AI Search, CosmosDB, Bicep IaC) for internal departments and external clients - adding multi-tenancy, MSAL auth, usage telemetry, and CI/CD on top of Microsoft’s open-source azure-search-openai-demo base.
- Delivered a crash-countermeasure benefit-cost tool over 9,777 FHWA studies across three platforms: a web app, a Power BI report, and a Tableau dashboard with a custom Extensions-API widget.
- Built the cloud side of a patent-pending IoT condition-monitoring system for stormwater trash interceptors: Azure IoT Hub ingestion, time-series data model, threshold alerting, and monitoring dashboard.
Selected projects
- Daycase - charting software for same-day surgery: scheduling, role-based patient charts, and case close-out. 29 clinical forms (about 500 questions), with clinical rules enforced on the server. TypeScript, React, Fastify, SQLite. Synthetic patients only.
Education
- B.S. in Computer Science, University of Miami, December 2023
Patent
- Named inventor, “Systems and Methods for IoT Sewer Management” - a load-cell system for condition-based maintenance of stormwater trash interception devices. Non-provisional application filed by Ardurra; patent pending.
Publications
Common Complications after Bariatric Surgery
Second author. Forthcoming in Laparoscopic, Endoscopic and Robotic Surgery, 2nd ed., 2027
From DEXA Data to Patient Action: AI-Assisted Body Composition Reporting in Bariatric Care
Sharma, P. "From DEXA Data to Patient Action: AI-Assisted Body Composition Reporting in Bariatric Care." Presented at the 7th International Bariatric Club World Congress, University of Oxford, September 2026. Abstract to appear in a British Journal of Surgery congress supplement.
RAGs in Data Pipelines Advance Water Industry Knowledge Consumption
Sharma, P., et al. "RAGs in Data Pipelines Advance Water Industry Knowledge Consumption." Presented at the APWAOK & OWEA Technical Conference, May 2025.
Skills
- Languages and frameworks: Python (pandas, FastAPI, SQLAlchemy), TypeScript/JavaScript (React, Vite), SQL (Postgres, DuckDB), HTML/CSS to production polish
- Data engineering: Parquet/S3 pipelines with incremental refresh, BI reverse-engineering (PBIX, TMDL, DAX), cross-vendor schema consolidation, migration validation and reconciliation
- AI engineering: guardrailed LLM pipelines in production (Claude API), RAG on Azure OpenAI + AI Search, agent tooling (custom skills, automated audits) - orchestration and product engineering, not model training
- Platform and domain: Azure (Static Web Apps, IoT Hub, Bicep, CI/CD), FHIR/Medplum and HIPAA-aware design, GIS (Leaflet, shapely), Docker, GitHub Actions