Tyler Locke
Learning Experience Designer · AI-Integrated E-Learning Developer · Instructional Technology Builder
Learning Experience Design

Practical learning, designed to be used.

Learning Experience Designer & Learning and Development Specialist

I design practical, interactive learning systems that turn complex workflows into clear, engaging training experiences.

I combine live facilitation, instructional design, process improvement, AI-assisted development, and front-end classroom tools to build learning resources that are easier to use, easier to scale, and easier to remember. The work below is a set of original sample projects and classroom-ready templates — built to show how I think and what I can produce.


Approach

Start from the problem

I don't build training to check a box. I find where learners actually get stuck, then design the smallest resource that fixes it.

Build

Make it usable

Job aids, classroom tools, and trackers a facilitator can open and run immediately — clean, low-friction, easy to trust.

Improve

Measure and tighten

I track readiness, scores, and follow-up so training keeps getting sharper instead of staying static.

About

About me

I came into Learning & Development the long way around. I started out teaching martial arts — up to a dozen classes a day, every age and every skill level — and that's where I learned the thing that still runs my work: people don't learn from what you tell them, they learn from what they practice. From there it was retail, mobile sales, and call centers, and in every one of those rooms I ended up being the person who trained everyone else.

What I actually bring to a team is an analytical mindset. The first thing I do with any document or process is look for how to make it better — but I won't propose a single change until I understand it completely. That order is the whole point: learn it inside and out, then improve it. It's why the things I build tend to hold up.

Across my career I've developed more than 4,000 people. At NYU Langone Health I've put that mindset to work across the entire program — I've revamped roughly 60% of it so far. And it's not just training material: I build walkthroughs, tip sheets, job aids, and process guides, plus countless trackers that tie multiple systems into one smooth workflow — Microsoft 365 apps, Power Automate, and Excel working together instead of sitting in silos.

I also just love to build. I've made my own AI assistant, Zareli, and it grows a little more every day — because I love learning, and building it is how I learn fastest. I bring that same curiosity to the AI I use at work: as a tool, with a human keeping it honest.

Underneath all of it, my approach is simple: solve the real problem. I don't build training to check a box. I find where people actually get stuck, design the smallest thing that fixes it, and keep everything clean enough that the next person can pick it up and trust it.

Selected Work

Sample projects & learning-design case studies

Original portfolio prototypes and classroom-ready templates — honest sample work, not employer or production deployments. Tap any project for the full case study, the file to open or download, and how to make it your own.

Handing these to another trainer? The maintenance walkthrough explains how to keep them updated.

AI in Practice

Using AI to build better training, faster

How I use AI in a learning-design workflow — with judgment and accuracy kept in human hands.

Where I stand. AI does not replace instructional judgment. I use it to speed up drafting, organize complex information, generate practice variations, and help trainers build better materials faster. Every output is reviewed for accuracy against the source of truth before a learner sees it.
Currently building Personal project · in active development

Zareli

An AI assistant and agent platform I'm designing and building from scratch.

Zareli (earlier working-named WAYLON) is a personal project where I'm building a full-stack AI assistant the same way I build training: solve the real problem, keep it clean, and keep a human in control. It pairs a Python/FastAPI backend with a React interface, and it's built around an approval-first model — the assistant proposes actions and a person confirms them, rather than acting on its own. It's the same "human in the loop" principle I apply to AI in learning design, turned into software.

  • Full-stack build — Python / FastAPI, React, local data store
  • Approval-first automation — actions are proposed, a human confirms
  • Two surfaces — an admin console and a user workspace
  • Designed for safety and traceability from the ground up

Where it fits in the workflow

Draft

Content development

First drafts of explanations, job aids, and outlines I then edit and verify.

Reuse

Prompt libraries

Saved prompts so any trainer gets consistent output without prompt skills.

Practice

Scenario generation

Variations of realistic situations so learners practice more than one example.

Assess

Knowledge checks

Draft questions and distractors from a workflow, reviewed for accuracy.

Clarify

Plain-language explanations

Readable rewrites of dense policy or system steps.

Coach

Coaching feedback drafts

Specific, constructive starters a coach can tailor.

Simplify

Workflow simplification

Long procedures compressed to the fewest steps that still work.

Verify

Human review

Nothing ships without a person checking it against the source of truth.

Sample trainer prompts

Starting points, not magic. Copy one, add your topic, then review the output.

Create a plain-language explanation

Explain [topic/process] to a brand-new [role] in plain language. Use one short analogy, keep it under 150 words, and end with one common mistake to avoid.

Generate roleplay scenarios

Write 3 short roleplay scenarios for [role] practicing [skill]. For each: the situation, what the other person says, and the outcome we want. Vary the difficulty.

Turn a workflow into a knowledge check

From these steps: [paste workflow]. Write 5 multiple-choice questions that test understanding, not memorization. Include one plausible wrong answer per question and mark the correct one.

Draft coaching feedback

A learner did this: [behavior]. Draft coaching feedback that names one specific strength, one specific area to improve, and one concrete next step. Keep the tone direct and supportive.

Build a facilitator guide

Create a facilitator guide for a [length] session on [topic]. Include objective, materials, timed agenda, key talking points, one activity, and debrief questions.

Simplify a policy into training language

Rewrite this policy for frontline staff: [paste policy]. Keep it accurate, remove legal jargon, and turn it into "what to do" steps. Flag anything that needs SME confirmation.
AI Toolkit

Working with AI

A practical page on getting real value from AI — how I pick a tool, prompts worth stealing, and the one habit that keeps AI useful instead of risky.

Picking the right AI for the job

No assistant wins at everything — anyone who says otherwise is selling something. Here's where Claude tends to lead for the work I do. The last row is where other tools have the edge, on purpose: leaving it out would undercut the whole point.

For this kind of work…
Why Claude is my default
Long documents & transcripts
Handles a very large context, so it can read and reason across long files, transcripts, and codebases in one pass.
Detailed, multi-step instructions
Follows complex instructions closely and holds a consistent voice across a long task.
Writing & rewriting
Clean, natural writing — strong for training content, emails, and documents that need the right tone.
Building something usable
Produces documents, slides, and working code you can preview and refine in place (Artifacts).
Showing its work
Tends to explain its reasoning and flag uncertainty instead of bluffing.
Where other tools lead
Real-time web depth, image and video generation, and voice. Pick the tool that fits the task — for building and writing, mine is Claude.

Confidence isn't accuracy.

Every AI — Claude included — can sound completely sure and still be wrong. For anything a learner or a patient-facing team will rely on, I treat AI output as a first draft and check it against the source of truth. Knowing that is the difference between using AI well and getting burned by it.

Switching to Claude without starting over

Coming from another assistant? You don't have to lose your context. Claude has a built-in import under Settings → Capabilities → Memory → Start Import: it hands you a ready-made prompt to run in ChatGPT, Gemini, or Copilot, that tool summarizes what it knows about you, and Claude pulls that profile in — so new chats start with your background already loaded. For a deeper transfer, export your full history from the other tool and have Claude work through the file.

Prompts worth stealing

Copy, paste into your AI of choice, and replace the [brackets]. The starred one is the one I use most.

Interview me to find my real need

You're helping me think this through. Before you answer, interview me — ask one question at a time to uncover what I'm really trying to accomplish, who it's for, my constraints, and what "done" looks like. Keep going until you genuinely understand, then play it back as a short summary of my real need and recommend the best approach.

Pressure-test it before I use it

Here's a draft: [paste]. Act as a sharp, skeptical reviewer. Point out anything that may be inaccurate, unclear, missing, or overstated, and give me a short checklist of what to verify before I rely on it.

Make it teachable

Take this [process or policy]: [paste]. Rewrite it as a plain-language explanation for a brand-new [role]. Use one short analogy, keep it under 150 words, and add a 3-question knowledge check with answers.

Turn rough notes into a clean draft

Here are my rough notes: [paste]. Organize them into a clean, logical [email / outline / one-pager]. Keep my meaning, tighten the wording, and flag anything that's unclear or seems missing.

Show me my blind spots

I'm about to [decision or plan]: [paste context]. What am I not considering? List the most likely risks, the assumptions I'm making, and the questions I should answer before I move forward.

Be my thinking partner, not a yes-man

I want honest pushback, not agreement. Here's my idea: [paste]. Tell me where it's weak, what I might be wrong about, and what a smart skeptic would say — then suggest how to make it stronger.

Explain it three ways

Explain [topic]: [paste] three times — first to a curious 10-year-old, then to a smart colleague from another field, then to an expert. I'll pick the level that fits my audience.

Give me the 80/20

Here's the full picture: [paste]. What's the 20% that gets me 80% of the result? List only the highest-leverage moves in priority order, and tell me what I can safely ignore for now.

Steelman the other side

I believe [position]: [paste my reasoning]. Now make the strongest possible case for the opposite view — the version a thoughtful person who disagrees with me would actually argue. Then tell me which points I can't easily answer.

Rewrite it at three lengths

Take this: [paste]. Give me three versions — a one-line summary, a short paragraph, and the full version tightened. Keep the meaning and tone consistent across all three.

Grade it against a rubric

Score this [draft/plan]: [paste] against these criteria: [list, or ask me for them]. Give each a 1–5 with one sentence of why, then rank the fixes that would raise the lowest scores fastest.

Reverse-outline what I wrote

Here's something I already wrote: [paste]. Work backwards and show me its outline — the main point of each section in one line. Then tell me where the structure drifts, repeats, or has a gap.

Role-play a hard conversation

I need to [have this conversation]: [context]. Play the other person realistically — including pushback. Go one exchange at a time so I can practice my responses, then afterward tell me what landed and what I could say better.

Turn this into a reusable template

I keep doing this task by hand: [paste example]. Turn it into a reusable, fill-in-the-blank template with [brackets] for the parts that change, plus a short checklist so it comes out consistent every time.
Meet

Meet Zareli — Tyler Locke's AI assistant

An AI assistant I'm building from the ground up — an approval-first co-pilot that thinks with me, not for me.

In active development · built by Tyler Locke

What Zareli is

Zareli isn't a skin over someone else's chatbot — it's my own platform, built from scratch around one rule: it proposes, I approve. A place to think, draft, organize, and automate the busywork, with a human always in the loop. I'm building it because the best way I've ever found to truly understand a tool is to build it myself — and then keep making it better.

How it's helped me so far

Thinking partner

A place to untangle messy problems

I talk a problem through and come out with structure instead of a tab graveyard.

Speed

Faster first drafts

Quick passes on content, outlines, and messages that I then sharpen — the blank page disappears.

Focus

One place, less switching

Fewer tools fighting for attention, more momentum on the work that matters.

Growth

It made me a better builder

Building Zareli is how I've learned FastAPI, React, and how AI systems actually fit together.

The plan

Here's where it is, and where it's going — honestly.

Now

A working assistant

Approval-first controls and a two-surface design: an admin console for me, a clean workspace for getting things done.

Next

Real, approved actions

Connecting the tools I use every day so Zareli can do things — once I sign off — not just talk about them.

The vision

A coordinated team of agents

Specialized agents that each do one thing well, working together toward a goal — with me conducting.

The vision

The “orchestra”

This is the part I'm most excited about — and I'll keep it in quotes, because it's where Zareli is headed, not where it is today. The goal is an “orchestra”: a set of specialized agents that each play one part well — research, drafting, checking, organizing, automating — coordinated so they work together instead of one model trying to do everything alone. I'm the conductor. Zareli keeps everyone in time, and keeps a human approving the big moves. I don't have the full orchestra yet. But I will.

Zareli logo displayed on a wall
Built quietly, on the side — because I love learning, and Zareli is how I learn fastest.

Zareli is a personal project, in active development — a working example of how I think about building with AI: ambitious vision, human in control, and accuracy that's earned, not assumed.

Visit Zareli.ai ↗
Documents & Links

Files, profiles, and notes

Resume, live portfolio links, and a few notes on how I work.

Resume

Resume (PDF)

Two-page resume — selectable text, ATS-readable. Opens in a new tab.

Live Site

Portfolio — Live Site

Add your hosted URL in portfolio/data.js

Live Site

Interactive Portfolio — Base44

Add your Base44 URL in portfolio/data.js

Notes

Learning-design notes

Solve the real problem, keep tools simple and trustworthy, design for recall, and measure readiness — not just attendance.

Notes

AI process notes

AI accelerates drafting and organizing. A human verifies accuracy against the source of truth before anything reaches a learner.

See the tools

Screenshots of the interactive training tools in this portfolio. Click any to open the live tool.

Supporting documents

Resume & Contact

Let's talk

Name
Tyler Locke
Email
lockeprocess@gmail.com
Location
Boynton Beach, FL

Available for Learning Experience Design, AI-integrated training, digital learning development, classroom tool design, facilitation support, and process-improvement opportunities.