Software Engineer @ blu Bank · Backend & Distributed Systems · AI Coding-Agent Infrastructure
Building reliable infrastructure for AI coding agents — agent runtimes, MCP servers, runtime guardrails, and developer tooling.
I'm a backend software engineer with 5+ years of experience, the last 2+ at blu Bank. My foundation is Java and Spring Boot; my focus is distributed systems — service boundaries, data consistency, and keeping production systems observable and operable.
Outside of work, I build infrastructure for AI coding agents. The work has moved in one direction — developer tooling and MCP servers that give agents real access to code and dependencies, runtime guardrails that enforce constraints, and multi-agent runtimes that make long-running workflows observable and reliable.
- agent-mesh — a runtime for persistent AI organizations: explicit authority, communication contracts, artifacts, scheduling, policy enforcement, and event-sourced state.
TypeScript - opencode-token-norm
— runtime guardrails for OpenCode: measures token/tool-call pressure, runs cost checkpoints, and provides controlled session handoff.
TypeScript
- Mechanism over prompting — rules a runtime can enforce beat rules a model is asked to remember.
- Measure, then claim — state sample sizes and limits; a result that turns out to be an artifact gets retracted, not kept.
- Explicit contracts — typed messages, schemas, and policy boundaries over implicit conventions.
- Operable by default — long-running systems should be observable, replayable, and fail loudly.
| Project | What it does |
|---|---|
maven-decoder-mcp  |
MCP server that lets coding agents inspect real Maven dependencies — decompile classes, search artifacts, and compare versions for breaking changes. Python · Java · MCP |
http2postman  |
Converts JetBrains / VS Code .http files into Postman Collection v2.1. One command, zero dependencies. Python |
| task-forge | Turns Trello cards into agent implementation jobs: git worktrees, Codex runs, PRs, board updates. Python |
| code-review-skill | A living, evidence-driven Agent Skill for high-signal code review. Agent Skills |
Distributed systems · Backend architecture · Event-driven systems · Developer tooling · AI agent infrastructure · MCP · Reliability · Observability
Java · Spring Boot · TypeScript · Python · Node.js · PostgreSQL · Redis · Docker