Case Study

PhysicianGPS is a provider search and analytics platform built for employers and organizations to help their members find the best physicians—not just the closest ones.

With healthcare quality varying widely, employees often struggle to identify top-performing doctors who accept their insurance, are nearby, and match their needs.

As the lead engineer on the project, I designed and implemented the end-to-end system architecture—from data ingestion and indexing to the backend APIs and frontend search experience.

The result was a scalable platform that enables sub-second search across 200K+ healthcare records, supported by an automated ingestion pipeline, reliable filtering, and a user experience optimized for speed and clarity.

When companies provide healthcare benefits, employees are left to navigate provider networks on their own. They often pick doctors based on convenience, proximity, or a friend's recommendation—not on objective quality or outcomes. This leads to frustration for patients and higher long-term healthcare costs for employers. The reality is that finding the right physician is difficult, even with modern insurance portals.

A healthcare startup, Avalon AI, aimed to address this issue. Their data science team developed machine learning models to evaluate physicians using billions of data points—from prescribing patterns and readmission rates to infection rates and outcome metrics.

To bring this to life, they needed an engineering team to build a fast, accurate, multi-tenant provider search experience on top of a massive dataset with strict tenant isolation for employer-specific networks.

PhysicianGPS became that solution: a platform that allows employees to quickly find high-quality, in-network physicians, while giving employers confidence that their employees are making data-driven healthcare decisions.

To address these challenges, I defined a set of functional and non-functional requirements that shaped the architecture and implementation strategy.

Building an effective provider search platform required more than just ingesting data and exposing it through a UI. The core challenge was understanding the complex and messy world of healthcare provider data.

Provider information comes from fragmented sources such as NPI registries, claims datasets, licensing boards, and insurance networks. Each source has its own structure, refresh cadence, and data quality issues. On top of that, high-value fields—such as specialties, hospital affiliations, outcome metrics, or telehealth availability—are not standardized and often require inference, cleaning, and enrichment.

At a high level, the platform follows a pipeline → index → API → UI model designed for speed, scalability, and tenant isolation.

To deliver a fast, trustworthy, and scalable experience, I implemented an architecture that cleanly separated data concerns, leveraged modern Next.js patterns for performance, and used AWS tooling to accelerate iteration and reliability.

Because PhysicianGPS was designed to handle sensitive data and power live decision-making, I built reliability and performance into the system from end to end.

The platform delivered measurable improvements for users, engineering, and the business.

PhysicianGPS reinforced an important lesson: the most impactful engineering happens when technical excellence and product clarity move in lockstep. Working across the full stack — from ingestion pipelines to search relevance to UI responsiveness — taught me the value of owning the problem, not just the code, and making decisions that balance speed, reliability, and usability.

PhysicianGPS pushed me to think beyond features and focus on building systems that truly empower people. By transforming messy healthcare data into a fast, trustworthy search experience, I helped create a product that gives users confidence in one of the most important decisions they make: choosing a physician.

This project strengthened my belief that great engineering is equal parts craftsmanship, curiosity, and accountability — that when we combine reliable architecture, observable systems, and thoughtful UX, we can deliver technology that feels effortless and makes a real difference. I'm proud of the platform I built, and I'm even more excited to take these lessons into the next challenge — solving complex problems, collaborating with sharp teams, and building products that matter.