Applied AI + Engineering Leadership

I figure out what people actually need, then I build it.

I’m Daniel. I lead Applied AI and automation work across enterprise systems, data, reporting, integrations, and internal tools. I use AI agents every day for research, prototyping, and implementation, and I stay involved in architecture and testing.

Leading AI + automation delivery at Southland Industries
BUILDING THE EXPERIENCEWorking product
01 · THE NEEDMake it fun to joinLive, visual, easy to use
02 · FIRST BUILDPrompt oneWorking round + architecture
03 · THE CHECK225 simulated participantsJoins, reconnects, scoring
ROOM D4G
ROUND 04Who’s this teammate?
~170 joinedREADY
Move through the field
Experience · 20+ yearsPublic-safety audio · ~$250K/year avoidedApplication modernization · 50% faster releasesAzure DevOps · 90% fewer deployment failures

Featured work

I wanted people to have fun.

I was leading a department meeting for about 170 people and needed an icebreaker. The tools I found could run polls and questions, but they could not create the shared experience I wanted, so I built it.

Name That Team Member leaderboard on the shared display
LIVE MEETING~170 people
Open the experience

Name That Team Member · 2026

A live team experience for the department meeting I was leading.

People joined from their phones while a host ran each round on a shared screen. The mystery image, reveal, scoring, and leaderboard gave the room something to enjoy together.

1prompt to a working round
~4prompts to the event-ready version
225simulated participants
Read how I directed the build

How AI helped

The first working round came from the first prompt. By the fourth, it was ready for the event.

The short prompt count came from giving the agent real engineering context: product intent, architecture, constraints, acceptance criteria, test expectations, and a clear milestone for each pass.

PROMPT 01 · WORKING ROUND

Start with one complete round.

I described the people, roles, product intent, architecture, constraints, acceptance criteria, and tests. The result was a complete working round.

Deterministic automation

I turned a repetitive animation task into a rules engine.

At ABCmouse, animators repeatedly assembled the same kinds of pronunciation and spelling segments. I encoded how letters, sounds, and surrounding context selected those components, saving hundreds of hours of repetitive work.

ABCmouse · Automated Animator

The same letter does not always make the same sound.

The tool analyzed a word, applied contextual sound rules, chose the matching animation components, and generated the repetitive sequence. This was deterministic automation, built before modern generative AI.

HUNDREDSof animator hours saved
See how the rules worked

TRY A WORD

WORDSHIP
LETTER / SOUND ANALYSISSH · I · P
CONTEXT RULES + H become one sound
ANIMATION COMPONENTS/sh/ + /i/ + /p/
FINAL SEQUENCE3 timed segments

Leadership

I still build, but I’m usually not building alone.

I set technical direction, lead solution and automation delivery, review approaches, coach developers and contractors, and stay close enough to the work to prototype, test, and help with the hard problems.

SOUTHLAND INDUSTRIES

Leading AI and automation delivery

I lead several active initiatives across enterprise systems, data, reporting, integrations, and internal tools, including automation work that supports the company’s Workday transition.

AI ENABLEMENT

Giving teams a reliable way to use agents

I write engineering guidance, coach contractors, review technical approaches, and define where automated tests can carry the work and where human judgment still has to decide.

HANDS-ON DIRECTION

Staying useful when the problem gets difficult

I prototype solutions, debug integrations, review architecture, pair with developers, and turn one successful implementation into patterns the team can reuse.

How I build with AI

I decide what the agent can prove and what still needs me.

Name That Team Member did not have a dedicated QA team. I directed an AI agent to create technical tests, exercise the system, and report defects while I kept responsibility for whether the experience felt clear, well paced, and fun.

That division is useful beyond one project. Agents can check joins, reconnects, synchronization, scoring, timing, browser behavior, regressions, load, and acceptance criteria. I decide what evidence is enough and whether the result actually solves the user’s problem.

See my GitHub
PLAN → BUILD

Give the agent enough context to make good decisions.

I define the problem, architecture, constraints, milestone, acceptance criteria, and evidence before asking for a working slice.

TEST

Use AI to cover the technical QA gap.

The agent builds automated checks and browser journeys, looks for regressions, and turns failures into a specific next pass.

REPEAT

Keep experience judgment human.

A passing test cannot tell me whether people will enjoy participating. I review the feel, pacing, clarity, and usefulness before release.

AI Lab

What I’m exploring.

I use small, working experiments to understand where a new tool is useful, what can be evaluated automatically, and where the limits show up.

MEDIA

Generative media

I’m testing image, video, and voice generation across hosted tools and local ComfyUI pipelines, with an emphasis on repeatable workflows and consistent output.

  • Image
  • Video
  • Voice
  • ComfyUI
EVAL

AI development and evaluation

I’m refining agent workflows, critic-and-builder loops, browser QA, automated evaluation, and the boundary between machine evidence and human review.

  • Agents
  • Gauntlet loops
  • Automated QA
  • Acceptance criteria
LIVE

Real-time systems

I’m continuing to work with WebSockets, authoritative state, reconnect behavior, and backend session patterns for reliable high-concurrency experiences.

  • WebSockets
  • State
  • Reconnects
  • Sessions

Experience

Twenty years of getting from idea to working software.

I’ve worked on early startup products, public-safety systems, and enterprise automation. Most of those jobs needed me to write software, make technical calls, and help other developers through the hard parts.

  1. 2025–NOW

    Lead Solutions Analyst

    Southland Industries

    I lead enterprise automation, data, and AI work, including several active initiatives and automation that supports the company’s Workday transition. I also set guidance for AI-assisted development and coach contractors.

  2. 2015–2025

    Software Engineer

    SDI Presence / Scientia Consulting Group

    I built and modernized enterprise and public-safety systems, including real-time audio, cloud services, and data workflows.

  3. 2012–2014

    Lead Software Engineer

    Age of Learning / ABCmouse

    I led a four-person team through a major move from Flash to JavaScript and built automation and standards that improved quality, production work, and onboarding.

Daniel Garcia
ONTARIO, CACurious by default.

About

I learn systems by making things with them.

I’m a software engineer who grew into leadership without wanting to give up the interesting parts of engineering.

Outside work, that usually means board games, game systems, small hardware builds, generative art, and an unreasonable Steam library. I like projects that give me a new set of rules to understand.

Computer Information Systems, University Coursework · Best Senior Project Award

Say hello

I’m always up for a good technical conversation.