Senior Software Engineer

Abraham Kang Senior Software Engineer (650) 661-2696 · [email protected] · Lake Arrowhead, CA · linkedin.com/in/abraham-kang-7a6381139/ Summary Senior Software Engineer with 10 years of experience building large-scale full-stack applications, cloud-native platforms, and distributed systems at Amazon and Google. Strong expertise in frontend and backend development using React, TypeScript, Java, Go, Python, and cloud platforms (AWS/GCP), with a focus on scalable, reliable enterprise solutions. Experienced in building AI-powered applications, including Agentic AI, LLM applications, multi-agent systems, and RAG architectures using LangGraph, LangChain, MCP, Amazon Bedrock, and Vertex AI. Professional Experience Amazon Senior Software Engineer Mar 2022–Present Architected a highly available, multi-region healthcare platform for Amazon Pharmacy by building scalable full-stack applications, AI-powered clinical automation, and secure cloud-native architectures with Agentic AI workflows, RAG pipelines, and HIPAA/FDAcompliant systems that reduced pharmacist workload by 30% across nationwide healthcare operations. • Led end-to-end development of an automated clinical verification platform for 24/7 virtual care, building a risk-based review UI with Next.js, TypeScript and Redux and designing scalable Node.js and Spring Boot services to enable real-time alerting, script comparison, and streamlined pharmacist intervention workflows. • Designed and implemented a real-time Drug Utilization Review Engine (DUR) using Java and Python, leveraging PyTorch-based risk models and AWS Glue with Amazon Aurora (PostgreSQL) to analyze patient medication histories and detect drug interactions, duplicate therapies, and dosage risks. • Implemented a low-latency regulatory compliance system using Go microservices to perform real-time prescription validation, utilizing Amazon DynamoDB for state-by-state rule retrieval and secure API integrations with State Boards of Pharmacy and NABP for license and PDMP compliance checks. • Engineered an LLM-driven clinical guideline summarization capability using Anthropic Claude models on Amazon Bedrock, building RAG pipelines with LangChain, Amazon Kendra, and Hugging Face NLP models to transform FDA/CDC protocols into reliable, sourcebacked clinical alerts while improving pharmacist decision support. • Developed the core orchestration layer for a Multi-Agent Clinical Customer Care Agent using Amazon Bedrock, LangGraph, MCPbased tool integrations, and FastAPI services, enabling intelligent routing, semantic search across SOPs, and secure access to backend order systems through DynamoDB-backed state management. • Implemented a scalable prescription processing and clinical event pipeline using Amazon Textract and Comprehend Medical, integrating Kafka-based messaging for asynchronous document workflows and orchestrating processing through AWS Step Functions with conditional human-in-the-loop verification. • Designed and implemented a secure, HIPAA/FDA-compliant clinical infrastructure with immutable audit logging using Amazon S3 Object Lock (WORM) and multi-AZ Amazon EKS deployments to achieve 99.99% availability for critical DUR services. • Integrated Amazon Macie into CI/CD pipelines to automatically scan Amazon S3 data stores for PHI exposure, triggering automated remediation and compliance alerts. • Mentored junior and mid-level engineers through PR reviews, design reviews, and career coaching focused on building scalable, compliance-aware healthcare systems. Google Software Engineer Jul 2016–Feb 2022 Software Engineer in the Google Ads Bidding & Auction team, building and maintaining large-scale ML-driven distributed systems with 99.99%+ availability and sub-100ms latency, supporting automated auction-time bidding, CI/CD pipelines, and reliabilityfocused infrastructure that contributed to high-scale revenue-critical ad serving. • Built large-scale ML bidding infrastructure leveraging TensorFlow on TPUs, TensorFlow Serving for low-latency inference, C++/Java for auction-time systems, and FlumeJava, MillWheel, Apache Beam, and BigQuery for high-throughput signal processing and simulation. • Engineered distributed data pipelines using Java, Go, Google Spanner, Mesa, MapReduce, and FlumeJava, alongside TFX (Python/C++) model training pipelines to enable continuous feedback and model retraining. • Developed internal auction and ML monitoring tools using TypeScript and Angular (migrated from Dart/AngularDart and Closure), integrating gRPC-Web for high-performance backend communication and leveraging Google Charts/Vizier for complex data visualization and experimentation insights. • Led authentication and authorization for Google Ads internal APIs, implementing OAuth2-based access control and service-to-service identity with gRPC, enforced via Java/C++ backend services and integrated with Google’s internal IAM infrastructure. • Engineered low-latency C++ ad serving infrastructure powering real-time auctions, implementing GSP-based auction execution, SIMDoptimized ranking, budget enforcement with Google Spanner, and large-scale ad retrieval using Bigtable. • Built external integration systems for Google Ads auction infrastructure, implementing AdX real-time bidding gateways in Go with gRPC/Protobuf, Auction Insights APIs in Java for serving competitiveness metrics, and real-time compliance filtering using Java/C++ and ML-based text classifiers. • Built and operated CI/CD workflows leveraging Piper-based source control and Rapid for automated canary-to-global deployments, implementing Monarch-based rollback automation and improving production reliability through blameless postmortems. Education Georgia Institute of Technology Aug 2014–May 2016 Bachelor of Science in Computer Science Skills AI / ML & Agentic Systems: LLM (Claude, Bedrock), LangGraph, LangChain, MCP, RAG Pipelines, Multi-Agent Systems, NLP, TensorFlow, PyTorch, TFX, Hugging Face, Vertex AI, Amazon Kendra, Comprehend Medical, Model Serving Languages: Java, Go (Golang), Python, C/C++, TypeScript, JavaScript, SQL Cloud & Infrastructure: AWS (EKS, DynamoDB, Aurora, S3, Glue, Step Functions, Bedrock, Textract, Macie), GCP (Spanner, Bigtable, BigQuery, Dataflow, GKE), Kubernetes, Docker, Terraform, CI/CD Backend & Distributed Systems: Microservices, Event-Driven Architecture, gRPC, REST, GraphQL, Node.js, Spring Boot, FastAPI, HighAvailability Design, Low-Latency Systems Data Engineering: Apache Beam, FlumeJava, MillWheel, Kafka, ETL/Streaming Pipelines, BigQuery, Spanner Frontend: React, Next.js, TypeScript, Redux, Angular Observability & Reliability: OpenTelemetry, Distributed Tracing, Production Debugging, Latency Analysis, Incident Response, 99.99% Availability Systems Security & Compliance: HIPAA, FDA, OAuth2, Encryption (AES/RSA), Secure Audit Logging (WORM), PHI Protection