by Marcus

Marcus Orochena

AI strategy lead, engineering leader, and fullstack builder with 15+ years shipping systems and teams.

[email protected]240-393-2034Remote · DC areaGitHub
Profile

Engineering leader and fullstack builder with 15+ years taking ambiguous problems to production. Currently AI Strategy Lead at The Phoenix: company-wide AI adoption plus the agent and product-intelligence systems teams use day to day. Earlier: scaled SaaS to multi-million users, led teams through acquisition, and founded products end-to-end.

The Phoenix

AI Strategy Lead Current

  • Own company-wide AI strategy across engineering, product, and ops—so teams get real leverage from agents while people stay accountable and human review stays the merge gate.
  • Ship an internal agent orchestrator (Elixir/Phoenix) that turns tickets and GitHub events into constrained runs—implementation PRs, reviews, and CI fixes—with tool allowlists and engineers as the merge gate.
  • Ship a product-intelligence workspace that grounds answers in live specs, indexed code, and design evidence—so decisions use citation-backed expected-vs-actual analysis instead of unconstrained generation.
  • Drove build-over-buy AI platform: a multi-provider gateway routes each workload to the best-fit model at API-token cost, keeps data in-house, and keeps engines swappable as the frontier moves.
  • Set adoption bands for AI-assisted work that baseline teams on leverage with ownership—not vanity usage—and drive enablement against those targets.

Software Engineer

  • Built an event management and scheduling platform serving 1M+ users across 100+ non-profits, plus an embeddable calendar widget (Next.js, Drizzle ORM, PostgreSQL).
  • Shipped a user-facing AI chatbot for personalized event recommendations using vector search, a custom MCP server, and multi-agent workflows.
  • Led a monorepo migration and owned DevOps across AWS and Cloudflare—CI/CD, networking, and deployment pipelines.

Sync or Swim ↗

Founder & Engineer Side project

  • Built a real-time bidirectional sync platform that keeps databases, CRMs, and vector stores aligned through low-latency change detection and propagation.
  • Designed extensible adapters for PostgreSQL, MySQL, Salesforce, Redis, and Vectorize—with one- and two-way sync, conflict resolution, retries, and a no-code schema mapper.

Enovational

Engineering Lead · Software Architect

Acquired by Ernst & Young in October 2022; continued as Manager, Software Architect through March 2023.

  • Led a development team on a SaaS platform serving ~3.3M users; doubled team velocity within three months through coaching, clearer goals, and better delivery practices.
  • Directly managed and mentored engineers through regular feedback, career development, code review, and hands-on technical guidance.
  • Owned AWS infrastructure and a stress-testing strategy that supported state-wide launches with zero downtime.
  • Built a public REST API and migration tooling; cut story points on portal integrations by ~50% through a clearer integration strategy.
  • Led post-acquisition infrastructure and tooling migration into EY with minimal user disruption.

Truck Driver Power

Senior Software Developer

  • Improved API response times 10× (~2000ms → ~200ms) and cut backend memory use 60% through query optimization, caching, async jobs, and GC tuning.
  • Set up CI/CD for web and mobile, moving release cycles from ~2 weeks to nightly, and containerized services for reproducible deploys.
  • Grew test coverage from 0% to ~90% and mentored six developers on TDD, git workflows, and code review.

Prudential Associates

Technology Officer · Founder, Threxys Technologies

  • Led document classification software for digital forensics and internal tools that automated time-tracking and billing.
  • Founded Threxys Technologies: sentiment and threat analysis for executive protection, plus cybersecurity consulting for small businesses.

Primary depth in Elixir, TypeScript, Rails, and AI systems. Comfortable ramping into unfamiliar stacks when the problem demands it—polyglot by practice, not by checklist.

AI systems
LLM agents, multi-agent workflows, MCP, RAG and vector search, AI gateways, prompt design, citation-grounded retrieval
Languages & frameworks
Elixir/Phoenix, TypeScript, Next.js, React, Ruby on Rails, Node.js, PostgreSQL, Redis, GraphQL, Drizzle ORM
Infrastructure
AWS, Cloudflare, Docker, CI/CD, ECS, Linux

Ongoing side consulting alongside full-time roles—architecture, product, and 0→1 builds.

Ernst & Young

  • Led an evidence-driven assessment of a mission-critical no-code platform serving 20+ state agencies—combining 15+ stakeholder interviews with codebase, production telemetry, and delivery analysis.
  • Produced an execution-ready investment roadmap spanning platform health, team operating model, 10 prioritized technical workstreams, and an incremental AI-native product capability.

Leyefe

  • Took two founders from concept to an iOS TestFlight beta within a few months, owning the mobile implementation and release.
  • Built a daily prompt-based photo network with rotating categories, media processing, location, notifications, social graphs, and a community feed.

Cosmop

  • Technical lead on a small founder-and-designer team; turned a vague concept into a production Rails application for internal beta, owning implementation and deployment.
  • Built travel discovery and curation around recommendations, maps, messaging, user-generated feeds, and community experts.