Houston Hoffman

Senior Software Engineer at NVIDIA (2017-10 – Present)

Led the design and rollout of NVIDIA's next-generation Remote Patient Monitoring (RPM) platform, enabling continuous biometric ingestion and AI-supported clinical triage for over 2M members across chronic and preventive care programs.
Led system design and technical architecture for large-scale Java-based and polyglot microservices, defining service boundaries, data models, scalability strategies, and fault-tolerance patterns for a platform serving millions of users.
Built clinician-facing dashboards and triage experiences using React, Next.js, TypeScript, Angular, and D3, significantly improving alert clarity, data exploration, and workflow efficiency for distributed care teams.
Developed and optimized a high-throughput, event-driven backend primarily in Java (Spring Boot) with Go and Node.js, leveraging gRPC, GraphQL Federation, and CQRS to process 3M+ clinical events/day with sub-150ms p95 latency.
Implemented an AI-driven triage system in Python/FastAPI, Django/Flask leveraging LangChain, OpenAI APIs, embeddings, and feature pipelines with guardrailed LLM orchestration on de-identified clinical data, reducing unnecessary ER escalations through real-time alert prioritization.
Designed and optimized relational and NoSQL data models using PostgreSQL, DynamoDB, Redis, and Aurora, including indexing strategies, partitioning, replication, and query optimization for high-volume transactional workloads.

Software Engineer at Qualcomm (2013-09 – 2017-10)

Played a key role in Qualcomm's transition from legacy platforms to cloud-native, microservice-based architectures, modernizing telehealth, Remote Patient Monitoring, and population health analytics systems.
Designed and applied a FHIR/HL7 clinical data ingestion pipeline using Kafka Connect on MSK, streaming into Redshift and BigQuery, reliably processing 600K+ clinical messages per day with strict FHIR resource validation and 99.9% delivery guarantees.
Increased system resiliency by redesigning retry, backoff, and failure-handling strategies across Kafka consumers and SQS-based workers, removing downstream contention and cutting burst-related processing delays by 60%.
Refreshed clinician-facing dashboards and workflows with React, JavaScript, TanStack Query, and MySQL, introducing shared UI components, real-time vitals visualizations, and rendering optimizations that reduced load times and drove higher care-team adoption.
Decomposed legacy C# into ASP.NET deployed on ECS Fargate, shrinking deployment cycles from hours to minutes while enabling elastic scaling for nationwide patient programs.
Built a real-time alert distribution pipeline using AWS SQS and Lambda (Python) with embedded ML-based filtering, lowering false-positive alerts and improving clinical triage response times.
Introduced a REST adapter layer to support analytics use cases, improving average API latency by 40% and enabling 50+ concurrent clinician dashboards without performance degradation.

Intern at Panasonic (2011-09 – 2012-10)

Built core learning flows and interactive lesson interfaces using PHP, CI, and JavaScript, improving UI responsiveness and reducing render overhead during peak usage.
Improved cross-platform UI components and patterns shared across Web, C and C++, enabling faster rollout of new learning modules.
Executed real-time lesson and chat UI using WebSockets and optimized state management, handling millions of daily events without degrading client performance.
Improved front-end performance by introducing code-splitting, memoization, and virtualization strategies, reducing initial load time and improving engagement metrics.
Automated end-to-end localization for web/mobile UI using Crowdin API and GitHub Actions, enabling rapid deployment of 40+ supported languages.

About