KYLEKILMETISOMEGA

Omega · Kyle Kilmetis · AI application building

I give ideas a working form.

I'm Kyle Kilmetis. I build applications across operations, commerce, mathematics, history, knowledge and games. I study the subject, find its underlying structure, and turn it into something another person can use, examine or experience.

AI is present in every project, but not always in the same place. Sometimes a model reads, compares, drafts, tags or recommends inside the application. Sometimes AI works beside me as I explore, specify, assemble and test a deterministic build. The purpose determines the role; human intention and judgment determine what becomes real.

Eleven interactive systems run on this page · start with any build ↓
Start with the end ↓
working builds, each with something you can operate
10working builds, each with something you can operate
server-enforced lifecycle states in one application
28server-enforced lifecycle states in one application
sheets in one mathematical drafting table; three operable here
27sheets in one mathematical drafting table; three operable here
generated pages, with every link and image checked
739generated pages, with every link and image checked

The subjects change. The work does not: understand the material, make its structure explicit, and build a form another person can operate. This page unfolds the way I build: the end first, then the parts in order, from a record that proves its facts to a Roman life judged by history, then the method that made them, and the return.

Everything here is my own work: I chose the subjects, worked out the rules, designed and built every application, every exhibit and this site. Where a project contains original written material, I wrote that too.

Prologue · The end, firstWhat is all of this for?

In conversation

Vision and mission

Two questions about what all of this is for. The answers are my own words; two more close the page.

  1. ASKED · 01

    What is your vision?

    KYLE

    The ascent of man through knowledge, discipline, creation, and transformation, carried now by a workforce that does not sleep. I see one unbroken line: cloud infrastructure becomes reliable platforms, platforms become agentic systems, agentic systems acquire memory, memory enables autonomous workflows, and autonomous workflows hand the one irreducible thing back to the human: creativity. The machinery exists to return us to purpose. I want intelligence built the way the ancients built their temples, to proportion, to purpose, and to last.

  2. ASKED · 02

    What is your mission statement?

    KYLE

    Unleash creation. Govern promotion. I build governed AI workforces, and the factories that produce them, so that the act of building becomes repeatable, verifiable, and owned by the people it serves. Nothing I ship is finished until something other than its builder says so.

The line this page follows

  1. 01 · PrologueThe end, firstWhat all of this is for
  2. 02 · Act IConsequenceRecords, bookings and advice that hold
  3. 03 · Act IIPrecisionSeats and figures proven exactly
  4. 04 · Act IIIMemoryIdeas and history joined to their evidence
  5. 05 · Act IVImaginationGods, a daily puzzle and a Roman life
  6. 06 · Act VThe methodThe cycle that made every one
  7. 07 · EpilogueThe returnWhere it is going, and your part

Acts I–IV · The buildsWhat does an idea look like once it works?

Working builds / Designed and built by me

Different subjects.
One way of making.

Eleven systems you can drive, every one designed and built by me. Choose a build, then run it: press, drag, step and replay. Each carries a blue Interactive marker saying how, and runs a rule from its build on synthetic data or the project's own published records; no client or customer data appears here.

I · Consequence

Workflows that carry consequences

Evidence, coordination, recommendations and human decisions.

II · Precision

Rules that must produce an exact answer

Geometry and mathematics whose results can be checked.

III · Memory

Knowledge that must keep its evidence

Sources, notes and written history made navigable.

IV · Imagination

Datasets made publishable and playable

One body of records turned into pages, puzzles and a game.

Act I · ConsequenceWhat has to hold when real work depends on it?

Bids, proposals & contracts

Prove the facts once.

An evidence-driven record for bids, proposals and contracts. Requests, assessments, scope, quotes and approvals are held in one connected record, so the same facts are not proven again at every step. The lifecycle is enforced on the server: 28 states, every transition audited, and AI output advisory only.

Shown here: lifecycle, evidence integrity, scope gate and quote reconciliation.

Designed and built by me, end to end.

Try it

Try the reconciliation step: quote lines are matched against the approved scope, and every difference needs a recorded decision before a proposal can be drafted. Synthetic line items; no client data.

Choose another build ↑

▶ Interactive · tap to tryAccept or challenge each quote line that differs from the approved scope; the gate opens only when every difference has a recorded decision.

LIFECYCLE · STEP 11 OF 17 · QUOTE REVIEW

LineScopeQuoteDecision
Site assessment, 1 dayMatched$960$960No decision needed
Materials, standard gradeMatched$520$520No decision needed
Labour, 3 daysVariance$1,140$760
Equipment hire, 2 weeksMissing from quote$1,450—
Expedited deliveryNot in scope—$384

Gate closed. 3 differences need a recorded decision before a proposal can be drafted.

Demonstration · synthetic line items · logic mirrors the application’s reconciliation rule

Functional overview · press play or step through the diagram

  1. 01

    Intake

    Request, site, client

  2. 02

    Assess

    Needs, observations, evidence

  3. 03

    Scope gate

    Verified evidence required

  4. 04

    Reconcile

    Every difference decided

  5. 05

    Propose

    Versioned, second approver

▶ Interactive · tap to tryPress Play to follow the path through the system, or step with the arrows and the step chips; each step lights the parts it uses.

CLIENTSWEBDOMAINSTORESField appoffline-first PWAOffice webreview · approveCRM (optional)adapter onlyNext.jstyped contractsFastAPI · domainLifecycle (28 states)Evidence integrityScope gateReconciliationProposals + approvalAudit trailAI gatewayadvisory only · redactedPostgreSQLrow-level securityEvidenceSHA-256, immutableOutboxqueue, retriesA request is not a commitment. Scope is not a quote. A recommendation is not an approval.Versions are append-only; each points to its predecessor.

Press Play to follow the path through the system, or pick a step.

Logical architecture, drawn from the application’s architecture notes · pattern level

Facts proven once. Next, time, which can only be spent once.

Act V · The methodHow does an idea become something you can operate?

How I build

From intention to evidence, and back again.

The tools change. The cycle does not: observation becomes interpretation, interpretation becomes a decision, and the decision takes a form that can produce evidence. What I learn returns to the next build as a better model.

  1. Observation
  2. Interpretation
  3. Decision
  4. Manifestation
  5. Evidence
  6. Reflection
  7. Improved model

  1. 01

    Capture the intention

    I begin with the thing someone should be able to do, understand or experience, and why it matters. The first expression can be incomplete; the work starts by turning it into a purpose clear enough to guide later decisions.

  2. 02

    Investigate the real situation

    I follow the work, material or experience as it actually exists. I look at the information, actions, handoffs, exceptions and unofficial compensations, while keeping observation separate from assumption.

  3. 03

    Find the underlying structure

    I identify the objects, relationships, states and rules beneath the surface. Similar structures recur across very different subjects, but the domain details determine whether the resulting application truly fits.

The other three stages ↓

  1. 04

    Derive the behavior and AI’s role

    I describe observable situations: who acts, what information is required, which conditions apply and what result should become available. Then I decide where interpretation or synthesis helps and where explicit logic should determine the behavior.

  2. 05

    Build a complete path and require evidence

    I prefer an early path that runs from a person’s action through the application’s rules to a visible result. The evidence must fit the risk: drawings, residuals, round trips, link checks, state transitions or another inspectable consequence of the behavior that matters.

  3. 06

    Carry the learning forward

    Use reveals what the original model missed. I preserve the reasons behind important decisions, the evidence that supported them and the assumptions that failed. Each project should strengthen the questions, methods and foundations available to the next.

Epilogue · The returnWhere is this going, and what must change for it to fit?

In conversation

The return

Two more questions: what has changed, and what has to change with it.

  1. ASKED · 03

    What excites you most about the current modern technological landscape?

    KYLE

    The bottleneck has moved. For the whole of my career the limit was hands: how many engineers, how many hours, how many files one person could hold in their head at once. That limit is gone. The limit now is intent and judgment: knowing what should exist, in what order, and what “done” means. Those were always the rare skills. Now they are the only ones that matter. And underneath that, something quieter and larger: memory is arriving. Systems that improve with experience, that carry the record forward, that learn the shape of a domain and keep it. One person with a method and a record can now run what used to take a department. A single man can command a legion.

  2. ASKED · 04

    What is it you mean by prioritizing the ‘Adaption’ of AI over the ‘Adoption’ of AI?

    KYLE

    Adoption is a procurement event. You license the tool, you bolt it to the side of the way you already work, you count the logins and call it transformation. Adaption is biological. It is the organism changing shape to fit a new environment, and the environment changing in return. When I say adaption I mean three things have to change at once: the AI is adapted to the work (its context, its tools, its limits), the work is adapted to the AI (what gets stated, recorded, and verified instead of assumed), and the people adapt most of all, in how they state intent and how they judge results. Adoption asks which tool. Adaption asks what must change in us for the whole to fit. The first produces demos. The second produces a new operating organism. I care about fitness, not ownership.

The line, completed

  1. ✓ PrologueThe end, firstWhat all of this is for
  2. ✓ Act IConsequenceRecords, bookings and advice that hold
  3. ✓ Act IIPrecisionSeats and figures proven exactly
  4. ✓ Act IIIMemoryIdeas and history joined to their evidence
  5. ✓ Act IVImaginationGods, a daily puzzle and a Roman life
  6. ✓ Act VThe methodThe cycle that made every one
  7. ✓ EpilogueThe returnWhere it is going, and your part

Epilogue · The returnWhat are you operating here, and how do we talk?

Talk with me

What you are—and are not—operating here

Each exhibit runs a real rule from its project using synthetic data or the project's own published records. This page does not expose client or customer data, credentials, private prompts, source code, or material that is not mine to publish.

If something here connects with an application, a role, a collaboration, a technical question or an idea you want to explore, email me.

Talk with meOperate another build
Who these are built with

Omega is my own portfolio. Some of the builds here are connected with EXARK AI: Agentic Proposal & Contract Desk, Scheduling, Intake & Outcomes Portal and The Wine Curator also appear, in their own copies, as demonstrations there. If one of them is relevant to you, ask me about it, or see exarkai.com ↗.

I keep my application work here; my engineering portfolio is separate: Alpha.