At BQE Software, I architected BQE Core AI Advisor, a production multi-agent platform built around supervisor/orchestrator and agent-as-tool patterns. I designed Agentic RAG with PostgreSQL/pgvector and Elasticsearch/OpenSearch hybrid retrieval, plus RAGAS pipelines for response accuracy, relevance, and faithfulness.
I build practical LLM workflows with AWS AgentCore, AWS Strands, Amazon Bedrock, Titan V2 embeddings, LangGraph, MCP, ReAct, FastAPI, and durable AI memory. My work includes event-driven embedding pipelines, HNSW vector search, and cloud-native deployment using AWS CDK, Lambda, ECS/Fargate, ECR, Docker, and Secrets Manager.
I also modernized legacy SQL Server workloads toward multi-tenant Aurora PostgreSQL through AWS DMS, CDC, logical replication, Kinesis, Kafka, SQS, and DynamoDB. Across BQE Core, I delivered REST and public APIs, SSO and OAuth flows, Hangfire workflows, Redis caching, Qrvey analytics, and application security hardening.
Previously at Omnicom Media Group, I integrated GPT-3.5 Turbo and DALLĀ·E 2 into Omni Assist for campaign insights, personalization, and creative ideation. I bring 9+ years of backend and AI engineering experience across Python, .NET, distributed systems, enterprise data, and production automation.
