Start with what needs to work
A useful architecture makes something easier to build, use, or operate. I start with the behavior people need and work backward to the system that supports it.
$2A · about the human
AI platforms by day. Arcade surgery after hours.
I'm Daniel Plas Rivera, a software architect in New York. I build AI platforms and data systems at work, then spend a lot of my own time figuring out how old arcade games run.
Today I work as a lead software architect at a global management consulting firm and hold an Anthropic-issued Claude Certified Associate — Foundations. My work includes agent tooling, model gateways, streaming interfaces, source-linked answers, classifiers, and evaluation.
Before that, I built web applications, APIs, data pipelines, and cloud infrastructure across financial services, higher education, commerce, and startups. I still enjoy moving through the stack to find the part that is actually causing the problem.
I also mentor engineers through code review, debugging, and delivery. A good handoff leaves someone able to change the system, not just keep it running.
Outside my professional work, I reverse engineer 1990s arcade software. That work spans 68000 and Z80 assembly, emulator cores, binary analysis, and a memorial fighting-game agent. Morphus56K is where those investigations meet my production engineering practice.
How I work
A few principles that travel between enterprise platforms and the arcade bench.
Make evidence part of the build
I build citation checks, evaluations, and runtime traces alongside the feature, so a failure leaves enough information to investigate it.
Keep the failure modes visible
I want to know what happens when a source is missing, a model is wrong, or a pipeline retries. Those cases belong in the design and the tests.
Leave a path for the next engineer
I write down decisions, explain the awkward parts in review, and help other engineers learn to debug and extend what we build.
Career context
The short version; the résumé carries the full chronology and technical detail.
2023 — present
Lead Software Architect
Global management consulting firm
Lead architecture and implementation for AI applications, from model access and retrieval to the APIs and interfaces people use. The work includes an agent tool-discovery layer, a multi-provider model gateway, source-linked responses, offline evaluations, and hierarchical classifiers.
2020 — 2023
Data platforms and reliability
Global management consulting firm
Led data engineering for multi-source analytics systems. Replaced repeated API polling with event streaming, built reusable schema checks and pipeline tests, and improved cloud security and operational tooling.
2019 — 2020
Contract software engineer
Regulated enterprise platforms
Built backend services for data-risk analysis and decision-support systems, with typed APIs, containerized delivery, and reviewable service boundaries.
2017 — 2019
Application architect and full-stack engineer
Financial services and cloud modernization
Defined target-state cloud architecture, rebuilt service boundaries, and improved delivery pipelines, observability, and runtime reliability.
2015 — 2017
Platform developer and engineering mentor
Higher education and developer learning
Modernized application stacks, connected legacy systems through APIs, and mentored engineers through project-based delivery and code review.
2009 — 2015
Full-stack lead and founder
Commerce, media, nonprofit, and startup work
Built and operated web products, led small delivery teams, and learned to translate unclear requests into testable software.
Behind the handle
Morpheus. A 56K modem. One stubborn misspelling.
Two references, both from the same era. Morphus is Morpheus from The Matrix, misspelled on purpose and kept that way — back then I wore my hair down to nothing and tinted shades, so the name stuck before I ever wrote a line of 68000. 56K is the modem. I went 14.4 to 28.8 to 56K, and at the time 56K felt impossibly fast.
The name is older than the work, but it aged well: most of what I do now is reading a machine's code until I can see what it actually does. The green on this site marks the same thing — addresses, hex indices, measured numbers, whatever the machine handed back.
The best way to reach me is email. You can also find my public work and professional history on GitHub and LinkedIn.