I’m a full-stack developer and AI engineer focused on bridging the gap between bleeding-edge AI models and robust, enterprise-scale software architecture. While the industry is flooded with fragile Python prototypes, my focus is production. I specialize in building highly scalable backends, using Java, Postgres, LangChain, and n8n to automate complex, multi-step business workflows that remain deterministic and predictable at scale.

My approach to agentic architecture focuses heavily on state synchronization, rigorous error-handling, data consistency, and human-in-the-loop guardrails. I design AI systems built to operate safely in real-world business environments — not just inside a terminal.

Experience

Aug 2023 — Present

Software Engineer · Epsilon

  • Architected a multi-agent code review platform on n8n — orchestrating four specialized agents (security, fraud, code quality, test coverage) with stateful session tracking and automated report generation, wired directly into the Azure DevOps CI/CD pipeline
  • Built an AI campaign-script automation pipeline with n8n, AWS Bedrock (Claude), and the Azure DevOps API, cutting script deployment from 60 minutes to under 2 — with human-in-the-loop approval gates and automated Git PR creation
  • Re-architected critical backend API endpoints, rewriting core logic and hardening exception handling to significantly cut response latency and eliminate a class of production errors
  • Delivered an end-to-end Private Label Credit Card (PLCC) payment integration that opened a new transaction path and directly grew revenue
Javan8nAWS BedrockPostgreSQLAzure DevOps

Feb 2023 — Aug 2023

Software Engineer Intern · Epsilon

  • Migrated the entire codebase from legacy Azure Repos to Git and stood up Azure Pipelines CI/CD with automated unit testing, sharply reducing manual-deployment errors and turnaround time
  • Built a customer referral program end-to-end in a .NET enterprise solution, driving measurable growth in referrals and downstream sales
.NETAzure PipelinesGitCI/CD
View full résumé

Projects

Xpanse.ai — Multi-Agent Campaign Intelligence

A production-grade marketing intelligence platform on LangGraph and Amazon Bedrock. Seven specialized agents run across two checkpointed pipelines — strategy generation and cultural content transcreation — grounded in RAG over dual Knowledge Bases, validated with live web search (Tavily), and gated by human-in-the-loop review at every strategic decision.

LangGraphAmazon BedrockRAGTavilyStreamlit
Read the case study
View full project archive

Writing

All posts →

Field notes on production AI — agent reliability, state management, and building enterprise backends that don’t break.

Hello World — My First PostJun 2026