Software engineer building production-grade AI systems and agentic platforms. I specialize in turning complex problems into scalable, reliable software — from orchestration and deployment to real-world integration.
Experience
AI Engineer
Travelers
Building AI-powered data intelligence platform using AWS Strands Agents and MCP. Evolved RAG-based SQL generation into sophisticated agentic solution for enterprise data analytics.
Co-Founder
Avyra
Co-founded healthcare data interoperability platform. Built multi-agent pipeline for automated PDF-to-FHIR conversion. Applied to Y Combinator Fall 2025.
Machine Learning Intern
Pitney Bowes
Built time series forecasting models using Prophet to predict partner transactional volume.
Data Science Intern
UnitedHealth Group
Optimized RDBMS structure for faster access and management of 1M+ customer records.
Featured Projects
Building AI systems that solve real problems with production-grade engineering.
Avyra (v2)
Retrospective Chart Extraction
The Problem: Clinical research teams spend weeks manually pulling variables from EHR records for retrospective studies.
Agentic tool that extracts clinical variables from Epic EHR data using FHIR APIs and Claude. Runs an autonomous extraction loop — fetching structured patient data, searching clinical documents, and returning results with source citations and confidence scores. Currently in talks with heads of orthopedics at Yale, Harvard, USC, and UMich about potential implementation.
Avyra (v1)
Healthcare Document Intelligence
The Problem: 80% of healthcare data is unstructured, costing hospitals thousands of hours in manual transcription.
Multi-agent pipeline that transforms unstructured medical documents into FHIR-compliant structured data. 6 specialized AI agents handle OCR, entity extraction, and medical coding.
Synced Ticketing Platform
Built for My Event Company
The Problem: Platforms like Eventbrite and Ticketmaster take a significant cut of every ticket sold — so I built my own.
I run an event company (@synced_official) and built this platform from scratch to cut out third-party ticketing fees. Handles tiered ticket sales with access codes, Stripe checkout, QR code generation, and real-time check-in scanning backed by Supabase.
Voxa
AI Voice Receptionist
The Problem: Small businesses miss calls and can't afford 24/7 reception staff.
Voice AI system that handles business phone calls using local LLM inference. Full speech-to-speech pipeline with privacy-first architecture.
NBA Salary Predictor
ML Classification Model
The Problem: NBA teams need data-driven approaches to evaluate fair market value during contract negotiations.
Machine learning model predicting NBA player salaries from performance statistics. Achieved 72% accuracy with tuned Random Forest classifier.
Skills & Technologies
Tools I use to build intelligent systems from concept to production.