// résumé
Ceryce Armstrong
AI Architect · Full-stack engineer of 10+ years
AI-focused software architect with over a decade of full-stack engineering experience, now specializing in agentic systems, LLM-powered developer tooling, and spec-driven AI development pipelines that have measurably accelerated delivery by up to 3×. Adept in architecture design, overseeing implementations, mentoring teams, designing multi-agent AI orchestration, maintaining strong client relationships, and acquiring deep domain expertise. Committed to translating emerging AI capabilities into reliable, production-grade systems that expand what one engineer can deliver.
Experience
Lead Engineer / AI Engineer · InRhythm, O’Fallon, IL
Jan 2026 – PresentJava 21 · Spring Boot · Spring Batch · Amazon EKS · Control-M · GitHub Copilot
- Designed and deployed a Spec Driven Development AI DLC pipeline to modernize legacy PL/SQL Oracle batch jobs running on the database host box to individual containerized Spring Batch jobs running in EKS. The legacy jobs were Control-M scheduled, and some remain so post-migration. This sped up development ~3× — a quarter per batch job down to a month.
- Generalized the above pipeline to modernize a Struts-based JSP legacy monolithic webapp into a modern Angular front end backed by domain-specific Spring Boot microservices. This improved development times by ~2×.
- Developed a reflexion memory system for the pipeline for it to learn from its mistakes and learn the client’s patterns and architecture better.
Founder · KMO Solutions Foundry
Apr 2025 – Jan 2026- Founded an independent web/hosting/integrations consultancy; designed and built its platform and portfolio, including an AI-native CRM, an AI call screener, and marketing/demo sites.
Senior Software Engineer · Bank of America, Chicago, IL
Apr 2024 – Apr 2025Java 17 · Spring · Spring Boot · DB2 · OpenShift · Kubernetes · Git · TDD · Kafka
- Contributing senior developer on a SAFe agile team of six developers, three onshore and three offshore (in India).
- Added new REST endpoints to a microservice for manipulating ACH transfer limits, vastly speeding up changes to them by eliminating the need for an entire application rebuild.
- Added resiliency in ACH fraud detection by adding circuit breakers to save resources and using the store-and-forward pattern with Kafka to store the transactions that were unable to process during downtime.
Senior Software Engineer · PNC Bank, Chicago, IL
Mar 2023 – Apr 2024Java 21 · Spring Boot · Spring Data · MySQL · Docker · Git · TDD
- Led a team of three building a greenfield Spring Boot middleware application in Java 21 that translated between legacy applications emitting batches and a new payment gateway accepting only REST calls, letting an entire fleet of legacy applications reach the new gateway without any of them being modernized.
- Designed it to be reconfigured rather than rewritten: chunk collection (configurable thread count and per-thread chunk sizing), batch-processing rules, and send logic are each independently swappable through Spring IoC, so a new batch format or a different downstream gateway reuses the collection-and-forwarding core instead of reimplementing it.
- Delivered under PNC DevOps standards — TDD, CI/CD on internal RHEL OpenShift, full monitoring/logging coverage.
Senior Software Engineer · Loop Commerce, Columbia, MO
Oct 2022 – Dec 2022- Loop Commerce was shut down by its parent company, Synchrony Bank, one month after I was hired.
Senior Software Engineer · SuprTek, O’Fallon, IL
May 2018 – Aug 2022Java 17 · Spring · JavaScript · Node.js · React · MongoDB · Git · TDD
- Led a team of eight converting a legacy, spreadsheet-driven logistics line of business as part of a full-scale USTRANSCOM application rebuild.
- Designed the data, service, REST API, and messaging layers for the Coronet line of business using MongoDB, Java 17, Spring Boot, and Kafka, and participated in the design and implementation of Coronet’s React-based UI.
- Served as interim Product Owner for six months, leading client interactions and advocating for the client in team decisions.
Senior Software Engineer · DMI, Carmel, IN
Nov 2017 – Mar 2018Docker Enterprise · Red Hat Linux · Bash · JSON · Artifactory · Jenkins · Microsoft Team Foundation Server · Maven · sed
- Managed the transition of Allison Transmission’s DevOps infrastructure from external DMI hosting to internal hosting.
- Migrated Jenkins, Team Foundation Server, Artifactory, and Git to Docker, building an image per service and moving the existing Maven artifacts into the new Artifactory.
Healthcare Informaticist · BJC Healthcare, St. Louis, MO
Nov 2014 – Jul 2017Java 6/7/8 · Spring · Struts · JavaScript · jQuery · Sybase · DB2 · SQL Server · PostgreSQL · Python · Linux · Bash · Git · CVS · Tomcat · MQ Series · Maven · Ant · Jira (administration and use)
- Led design and implementation of BJC’s CDC grant-funded software solution for antibiotic use and resistance reporting.
- Led conversion of major projects’ build infrastructure to Maven, mentoring six team members on its use.
- Maintained and updated hospital-acquired disease reporting infrastructure using Java (6, 7, & 8), Spring, Struts, JavaScript, and jQuery.
- Interpreted CDC technical guides to update and implement new XML reports for CDC submission.
Software Associate · National Information Solutions Cooperative, St. Louis, MO
Jun 2014 – Nov 2014Java 5 → 8 · Swing · JDBC · SQL · Oracle Database
- Supported the CIS and Cash Register Java applications, focused on adding new UI pages to Cash Register.
- Improved the performance of a SQL query by over 1,000× in relevant edge cases.
Skills
- Java (6–21)
- Lua
- C#
- C++
- JavaScript
- Python
- Bash
- PowerShell
- Lisp
- Spring
- Spring Boot
- Spring Data
- JUnit
- Mockito
- Node.js
- MySQL
- PostgreSQL
- MongoDB
- MSSQL
- Oracle
- DB2
- Hibernate
- Claude Code
- Claude Agent SDK (Py & TS)
- GitHub Copilot
- Cursor
- Prompt engineering
- Agentic orchestration
- Subagent design
- Spec-driven development
- Skill & hook dev
- MCP servers
- Docker
- Tomcat
- Caddy
- Linux
- Gradle
- Maven
- Ant
- OAuth2
- Kubernetes
- OpenShift
- Kafka
- IBM MQ
Certifications
Selected projects
Full portfolio ↗- Demiurge — meta-agent lifecycle systemA Thrallherd-pattern meta-agent that mints, specifies, deploys, and curates a roster of specialized, long-running “Archon” agents — owning the full spec → eval-gated admission → delegation → tenure/retirement loop. Archons run as subscription-billed CLI services behind an A2A/MCP interface; the first production Archon autonomously fetches and scores job postings, researches companies, and tailors application materials under a draft-and-hold approval gate.
- Margo — agentic chief-of-staffA persistent, local-first AI orchestrator: purpose-built subagents loaded just-in-time, proactive loops that screen calls, texts, and email and manage daily scheduling, and an “act-low / ask-high” gate that drafts-and-holds anything outbound. Observability — append-only per-turn telemetry (verbatim turns, tool calls, tokens/cost/latency) joined on one turn id and stamped with the vendor CLI version, which attributed a 40% latency / 45% cost drop across 3,000 turns to a Claude Code upgrade, not an internal change, in twenty minutes, no A/B. Security — no secret enters argv, logs, or the agent transcript: Worker secrets piped to wrangler over stdin (9 tests, one asserting the value appears nowhere in output), and a recipient-class ledger at all 6 outbound send sites that logs who was reached, never the address (51 tests). Governance — merge approvals single-use, bound to one head SHA, 24 h expiry, enforced by a fail-closed pre-tool hook; 58 owner rulings in a CI-checked ledger, 28 enforced in code; 6 blocking CI policy checks.
- Spec-driven AI development pipelineA self-verifying, spec-first engineering pipeline with a durable task-graph runtime and typed contracts, using a dedicated verification gate so the agent doing the work never grades its own output — plus a reusable suite of 7 Claude Code skills, including context-loaders and crash recovery, to make multi-agent engineering repeatable.
- AI-native CRM & call screenerA multi-tenant customer-engagement platform — architecture, data model, and specifications mine, implementation directed through coding agents rather than hand-written — with an AI intake pipeline that screens live calls through an LLM persona that reads caller intent, transcribes voicemails into structured leads, triages submitted photos, and drafts review replies.
- Distributed agent fleet & commodity-market automationA distributed control plane for a large fleet of autonomous client agents — a Java/Spring/PostgreSQL server tracking and commanding Lua-scripted clients in real time, plus a monitoring and command-and-control web app for fleet-wide visibility and operator intervention — and an automated commodity-market trading system that priced and executed trades across a large-scale virtual economy, with a 1,700+ function reference manual and instructional video series to onboard other operators.
Education
B.S. in Computer Science
Aug 2010 – Dec 2014South Dakota School of Mines & Technology — Rapid City, SD
Get in touch: contact@ceryce.com