Looking for a job

Stephen Chen

@stephenchen1

Message

Senior software engineer specializing in RAG retrieval, grounded LLM systems, and safety-critical healthcare AI.

United States

Message

What I'm looking for

I’m looking for a team where safety-critical correctness matters—shipping grounded RAG/LLM features with rigorous evaluation, strong observability, and regulated reliability. I want to own retrieval quality, gating, and performance (latency/cost), not just prompts.

I’m a senior software engineer with 10+ years building backend platforms where correctness and reliability directly impact outcomes. At Viz.ai, I lead the retrieval-and-serving layer for LLM features over clinical data, chunking and embedding FHIR records into a retrieval index so an LLM (Amazon Bedrock) can answer and summarize with every output cited back to source for clinician trust.

I own the parts that decide whether an AI feature can ship: whether vector search and re-ranking surface the right context, whether the answer is grounded in source, and whether the LLM evaluation harness (grounding and hallucination checks) holds the quality bar under real latency and cost. I lifted retrieval recall@10 by ~18 points (72% to 90%) with a cross-encoder re-ranker, stood up an evaluation harness that cut the unsupported-claim rate under 3%, and built real-time ingestion and alerting pipelines (Python microservices on AWS EKS with Temporal and RabbitMQ) so relevant patient history rides along with each flagged study—built and operated under HIPAA in a SOC 2 Type II environment.

Experience

Work history, roles, and key accomplishments

Current

Senior Software Engineer

Current

Viz.ai

Jan 2023 - Present (3 years 6 months)

Led the retrieval-and-serving layer for LLM features over clinical data, chunking and embedding FHIR records and generating cited answers via Amazon Bedrock. Improved retrieval quality with an end-to-end pipeline and re-ranker, and built an LLM evaluation harness to gate prompt/model changes using grounding and hallucination checks.

Retrieval Augmented Generation (RAG)Vector Search Embeddings Cross Encoder Re Ranking LLM Evaluation FHIR Python LLM As A Judge RabbitMQ Microservices Prometheus Grafana OpenTelemetry HIPAA

Software Engineer II

Q2 Holdings

Sep 2019 - Nov 2022 (3 years 2 months)

Built the Kafka event backbone to stream account activity into Q2 Sentinel for real-time fraud scoring, reducing fraud-signal latency from overnight batch to under ~5 seconds. Migrated services from .NET Framework to .NET Core, delivered white-label UI theming with Angular/React/TypeScript, and built deployment tooling for a per-financial-institution model.

Microservices Event Streaming C .NET Framework .NET Core Linux Docker Angular React TypeScript Nomad Azure DevOps

Software Engineer

Microsoft

Oct 2016 - Mar 2019 (2 years 5 months)

Built Azure Service Fabric backend services and ASP.NET Core control-plane REST APIs for an Azure platform team, with stateless/stateful reliability patterns and service-to-service authentication. Authored ARM templates and CI/CD wiring, implemented telemetry/DRI on-call, and shipped a ring-based safe deployment approach with an Angular/TypeScript admin UI.

Azure Service Fabric Reliable Collections ASP.NET Core ADAL Application Insights Geneva (Internal Telemetry Stack)Ring Based Deployments Angular TypeScript

Software Engineer

LawnStarter

Jun 2014 - Sep 2016 (2 years 3 months)

Built marketplace features on a PHP/Laravel backend with MySQL, including customer signup as an AngularJS single-page app and improvements to signup completion. Implemented Stripe payments and weekly provider payouts with Google Maps Places/geocoding integration, and helped ship iOS and Android provider/customer apps.