DEV Community: Delon Swartz

AI Product Design Practice: Agents, Interfaces, and the Craft of Shipping

Delon Swartz — Fri, 25 Sep 2026 14:47:14 +0000

Meta description: A profile of Delon Swartz on AI product design practice — agent UX, shipping discipline, and building AI-native products from South Africa without the hype cycle.

AI product design is not “UI for chatbots.” It is the discipline of turning probabilistic systems into reliable products: clearer intent, safer actions, measurable outcomes, and interfaces that respect human attention.

Delon Swartz is an AI Design Engineer and product builder in Durban, South Africa. He works across Applied AI, Applied ML, and AI product design; founded ModusMax AI; and ships end-to-end products through Delon Labs UX. This essay shares his practice — how he thinks about agents, UX, and delivery — without fabricating portfolios or vanity awards.

The object of design has changed

Classical digital product design assumed deterministic software. Click → known state. AI-native products assume:

Outputs vary.
Tools can take actions with side effects.
Context windows and memory change behaviour over time.
Users must negotiate control with the system.

That means AI product designers (and hybrid AI design engineers) inherit concerns that used to sit only with SRE, security, and ML ops: permissions, observability, evaluation, and graceful degradation.

If a design system has a button component but no pattern for “agent is about to modify a file — confirm?”, it is not yet an AI design system.

Practice pillar 1: Intent over prompt theatre

Users should not have to become prompt engineers to complete a job. Good AI product design:

Offers structured intents (“Summarise this thread for legal review”) alongside free text.
Captures constraints in UI (tone, length, data sources, deadline).
Turns successful prompts into reusable skills or playbooks.

In agent platforms, this shows up as skills, spaces, and clear task framing rather than a blank box and hope. ModusMax’s product direction — desktop-first agents, local or cloud engines, chat plus operational surfaces — is built around that belief: power without terminal gatekeeping.

Practice pillar 2: Agency with brakes

“Agent” means the system can plan and act. Product design must decide how much agency is unlocked at each trust level.

A practical ladder:

Suggest — draft only.
Assist — act inside a sandbox.
Execute with confirm — propose a tool call; human approves.
Execute within policy — autonomy inside an allowlist.
Supervise — human reviews exceptions and audits.

Climbing this ladder too fast is how demos become incidents. Climbing too slowly is how products feel useless. The craft is matching ladder height to domain risk — a theme that runs through Swartz’s agent UX work.

Practice pillar 3: Make the machine’s mind legible (enough)

Full transparency of every token is rarely usable. Strategic legibility is:

What sources were used?
What tools were called?
What is the plan before irreversible steps?
How does the user stop, undo, or take over?

Logs and timelines are UX. So are empty states that teach mental models (“Can read these folders; cannot send email until Gmail is connected”).

Practice pillar 4: Design the boring path

AI product marketing loves the wow path. Real users live on the boring path: login, permissions, partial failures, unknown/uncertain answers, rate limits, offline mode, and handoff to a human.

Swartz treats the boring path as first-class:

Explicit failure copy.
Partial success states.
Export and escape hatches.
Support surfaces that do not require the user to screenshot a raw stack trace.

This mindset comes as much from shipping full-stack products (UX through backend and deploy) as from model work. AI features that cannot be operated will not be loved.

Practice pillar 5: Build with AI, design with judgement

Modern builders use AI-assisted development tools. Swartz does too — they compress the distance from concept to working software. That speed is valuable. It is also dangerous if judgement does not keep up.

AI-assisted coding does not absolve product ethics, accessibility, security reviews, or taste. His practice is: move faster on implementation, stay slower on claims. Especially in public writing, he prefers precise statements over inflated ones. Markets punish fiction.

South African vantage point

Designing AI products from South Africa is not a footnote. It is a vantage point that shapes his work:

Global tooling, local users — ship with world-class models and frameworks while designing for African bandwidth, pricing sensitivity, and mobile habits where relevant.
Founder-scale ownership — many SA builders wear design, ML integration, and delivery hats in one role. That hybrid is exactly where AI Design Engineering lives.
Narrative responsibility — the market does not need imported hype cycles. It needs working agents, honest UX, and products like job platforms that serve real people (for example, youth-focused tools such as TheJobPlug).

He has also shared public concept work — from streaming co-viewing UX explorations to notes on video agents — because product practice improves when ideas meet critique early.

A weekly rhythm that keeps practice honest

For individuals and small teams, a lightweight rhythm beats a heavy process cult:

Cadence	Focus
Daily	Ship a thin vertical or kill a bad assumption
Weekly	Review eval samples + 3 real user/session notes
Bi-weekly	Permission and safety review for new agent tools
Monthly	Re-read positioning: are we solving a job or chasing a demo?

Write in public when possible. A clear LinkedIn article or Dev.to post forces sharper thinking than an internal wiki nobody reads — a habit Swartz models in his own posts.

What he is building toward

ModusMax is his clearest statement of AI product design belief: agents should be operable by people who do not live in a terminal, with local and cloud options, and product surfaces that make agency understandable. Early access details are on modusmax.online.

Alongside that, Delon Labs UX continues the habit of designing and shipping complete products — because AI product design that never touches production is only theatre.

Connect

For product, design, or Applied AI leaders seeking a practitioner conversation — South Africa–based, globally minded:

AI product design will keep changing with every model release. The durable skills are older: empathy for users, respect for risk, and the discipline to ship.

About

Delon Swartz is an AI Design Engineer in Durban, South Africa, focused on Applied AI, Applied ML, and AI product design. He is the founder of ModusMax AI and builds full-lifecycle products via Delon Labs UX. Connect on LinkedIn.

Applied ML in Corporate Product Design: From Model Demos to Dependable Experiences

Delon Swartz — Fri, 25 Sep 2026 14:39:41 +0000

Meta description: A practical guide to Applied Machine Learning inside corporate product design — workflows, UX, risk, and evaluation — featuring Delon Swartz, AI Design Engineer in South Africa.

Every corporate roadmap now has a machine learning slide. Fewer roadmaps have a clear answer to a harder question: how does Applied ML change the product design practice itself?

Delon Swartz is an AI Design Engineer working from Durban, South Africa, with a public focus on Applied AI, Applied ML, agentic workflows, and shipping complete products. He founded ModusMax AI and builds through Delon Labs UX. This article is a field guide — evergreen, practical, and free of invented case metrics — drawn from how he approaches Applied ML so features hold up in production.

The gap between a demo and a designed product

A notebook accuracy score is not a product. A chatbot that answers once in a boardroom is not a product. Corporate Applied ML succeeds when three systems agree:

The learning system — data, training or prompting strategy, evaluation, monitoring.
The interaction system — UI/UX, permissions, explanations, recovery.
The operating system of the business — policies, roles, SLAs, audit, change management.

Designers who ignore (1) decorate risk. ML engineers who ignore (2) ship power tools without handles. Leaders who ignore (3) create shadow IT with a neural network attached.

Applied ML in product design is the craft of aligning all three.

Design principles for Applied ML features

1. Probabilistic UX is still UX

ML outputs are uncertain. Interfaces must say so without drowning users in jargon.

Useful patterns include:

Confidence-aware presentation (when calibrated confidence is available).
Citations or source panels for retrieval-based answers.
“Suggest, don’t silently mutate” defaults for high-impact actions.
Easy correction loops that improve the system or at least the immediate task.

In agent-style products — including the direction Swartz takes with ModusMax, where agents act across desktop and cloud contexts — the design bar is higher: actions can touch files, browsers, and external tools. Permissioning and visibility are not polish; they are the product.

2. Start with decision rights

Before model selection, decide:

Who may approve an automated action?
What is reversible vs irreversible?
Which data classes are in-bounds?
What happens when the model is wrong at 02:00?

These are design decisions as much as compliance decisions. They belong in flows and empty states, not only in a policy PDF.

3. Prefer thin slices with thick evaluation

Corporate teams often over-scope “AI transformation.” A thinner vertical — one workflow, clear success metric, solid eval harness — beats a platform fantasy.

Evaluation should mix:

Offline: golden sets, regression tests, safety checks.
Online: task success, edit distance / correction rate, latency, abandonment.
Qualitative: trust interviews with the actual operators, not only executives.

If a team cannot describe how it will know the feature is worse next month, it is not ready to ship.

4. Data pipelines are part of the experience

Users feel pipeline pain as “the AI is dumb today.” Freshness, access control, schema drift, and feedback capture are product concerns. Swartz’s skill focus areas — Applied AI, ML, data pipelines, and agentic workflows — exist because product quality collapses when data plumbing is an afterthought.

Corporate contexts where Applied ML design earns its keep

Without inventing confidential client stories, the types of problems that respond well to careful Applied ML design include:

Knowledge work copilots that retrieve from approved corpora with audit trails.
Operations assistants that draft, classify, or route — with human confirmation on edge cases.
Customer experience aids that summarise and suggest, while keeping brand and regulatory voice under control.
Internal agent platforms that reduce repetitive desktop work without forcing every employee to learn a terminal.

South African corporates share these needs with global peers, with extra sensitivity to cost, connectivity, and multilingual realities. Designing for those constraints is a competitive advantage, not a limitation.

A simple delivery blueprint

When Swartz partners with teams or builds his own products, the sequence looks like this:

Frame the job — primary user, frequency, cost of error.
Map the current workflow — where judgement sits today.
Prototype the interaction, not only the model — clickable or Wizard-of-Oz if needed.
Select the ML approach — classical ML, embeddings + retrieval, LLM prompting, fine-tuning, or tools/agents — based on the job.
Instrument — logs that respect privacy but enable learning.
Pilot with champions — small cohort, real data, real stakes.
Harden — latency budgets, fallbacks, access control, runbooks.
Narrate change — training and UX copy that set accurate expectations.

This is deliberately unglamorous. Glamour is how demos win meetings. Dependability is how products win years.

What South Africa adds to the conversation

Working from Durban and across South African networks influences how he thinks about Applied ML design:

Build for mixed digital maturity inside one organisation.
Respect mobile-first and WhatsApp-shaped habits when customer touchpoints matter.
Design for teams that must justify ROI quickly — Applied ML should read as operational improvement, not research theatre.
Participate in the global tooling wave — cloud AI, local models, AI-assisted development — while keeping ownership of product judgement.

He shares concepts and research publicly on LinkedIn for the same reason he encourages teams to share carefully: closed laptops do not create markets. Open, precise thinking does.

Avoiding two common failure modes

Failure mode A: Model-first theatre.

A vendor model is selected before the workflow is understood. Design is asked to “make it pretty.” Users never adopt it.

Failure mode B: Design-first fiction.

Beautiful screens promise magic the data and models cannot support. Trust collapses on first contact.

The antidote is a shared language between design, ML, and engineering — and practitioners who can walk between those rooms. That is the niche of the AI Design Engineer / AI product engineer: not claiming omniscience, but translating.

An invitation, not a brochure

This article will not pad with fake before/after percentages or unnamed Fortune logos. For working artefacts, look at the ModusMax product direction at modusmax.online, public UX/AI writing on LinkedIn, and shipped product work such as TheJobPlug.

Organisations moving from ML demos to dependable product experiences — in South Africa or with distributed teams — will find in Swartz’s work a clear example of Applied ML treated as product design, not research theatre.

Applied ML does not need more mystique in the enterprise. It needs product design that tells the truth about uncertainty — and still helps people finish the job.

About

Delon Swartz is a Durban-based AI Design Engineer focused on Applied AI, Applied ML, and AI product design. Founder of ModusMax AI and product builder at Delon Labs UX. LinkedIn · ModusMax.

Applied AI Design Engineering in South Africa: Building Products People Actually Use

Delon Swartz — Fri, 25 Sep 2026 14:38:05 +0000

Meta description: How Applied AI, thoughtful UX, and product craft meet in South Africa — featuring Delon Swartz, an AI Design Engineer working at the intersection of agents, ML, and product design.

South Africa’s technology scene has always punched above its weight: deep engineering talent, sharp product instincts forged in constrained markets, and a bias toward shipping. In that environment, Applied AI is no longer a slide in a strategy deck. It is a design and delivery discipline — how models, data, interfaces, and workflows come together so people get work done.

This piece is about that discipline. It also introduces the approach of Delon Swartz, an AI Design Engineer and AI product practitioner based in Durban, KwaZulu-Natal. Publicly he also describes his work as AI Product Engineer & Automation Specialist, spanning Applied AI, machine learning workflows, agentic systems, and product UX. He founded ModusMax AI, a desktop-first AI agent platform, and builds digital products end-to-end through Delon Labs UX.

What follows is not a fabricated case-study reel. It is a practical view of Applied AI design from a practitioner shipping in South Africa — and a window into how that craft looks when design, ML, and delivery sit in one role.

What “Applied AI” means when you also design the product

Applied AI is often split from “design”: engineers train or wire models; designers polish screens. That split is outdated.

In practice, Applied AI design engineering means:

Choosing the right level of intelligence — when a rules engine, a retrieval layer, a fine-tuned model, or a multi-step agent is the honest answer.
Designing the human loop — confirmations, overrides, explanations, and failure states that keep trust intact.
Owning the path to production — auth, APIs, data pipelines, evaluation, and deployment — not only a prototype.
Respecting local context — bandwidth, device mix, languages, and business processes that look different in Johannesburg, Cape Town, Durban, and beyond than in a US demo video.

If the model is impressive but the workflow is confusing, the product has failed. If the UI is beautiful but the agent hallucinates irreversible actions, the product has failed harder.

South Africa as a proving ground

Working from South Africa shapes Applied AI work in useful ways — and Swartz’s practice reflects that environment.

Constraint breeds clarity. Teams here rarely have infinite GPU budgets or endless specialised headcount. That pushes toward pragmatic stacks: strong cloud services where they earn their keep, local or lighter models where privacy and cost matter, and ruthless prioritisation of user jobs-to-be-done.

Corporate and startup demand are rising together. Banks, insurers, retailers, media, and public-interest organisations are exploring automation and copilots. At the same time, founders are building agent platforms and vertical tools for African and global users. The shared need is people who can translate capability into product.

Trust is the product. South African users — like users everywhere — abandon tools that feel opaque or unsafe. Applied AI design has to make risk visible: what the system will do, what it will not do, and how a human stays in control.

None of this requires hype rankings or invented awards. It requires craft.

Where his work sits today

A few publicly shareable anchors:

ModusMax AI — Swartz founded and builds ModusMax, a plug-and-play AI agent experience delivered as a desktop app (macOS and Windows), with cloud continuity and options such as Telegram access and local Llama models. The product thesis is simple: powerful agents should not require a terminal ritual. Early access and product details live at modusmax.online.
Full-lifecycle product building — Through Delon Labs UX he designs and ships complete digital products: UX, architecture, backends, AI integration, and deployment, often with modern AI-assisted development tooling.
Shipped consumer-facing work — Including TheJobPlug (thejobplug.co.za), a South African job/CV platform aimed at youth entering the market.
Earlier industry grounding — Production coordination experience at Triggerfish in Cape Town on Netflix-related animation work, which sharpened delivery discipline long before agents were mainstream.
Ongoing public thinking — LinkedIn essays and concept shares on UX research, co-viewing product design, agentic AI, and emerging ideas such as video agents.

He also continues to deepen formal Applied AI skills (including Microsoft generative AI engineering coursework listed on his profile). Titles evolve; the through-line is consistent: design + AI systems + shipping.

A working framework for Applied AI design

When Swartz approaches a corporate or startup AI product problem, he uses a simple loop:

1. Job, not model

Start with the user’s job and the organisation’s risk profile. “Add GPT” is not a requirement. “Reduce time-to-quote without increasing compliance risk” is.

2. Interface before spectacle

Map the moments where AI appears: empty states, progressive disclosure, citations, undo, escalation to a human. Agents that act on files, browsers, or customer records need explicit permission patterns.

3. Evaluation as a product feature

Offline metrics matter. So do online ones: task completion, correction rate, time saved, and qualitative trust. Design instrumentation early.

4. Architecture that matches maturity

Many teams need retrieval-augmented generation and solid tooling before multi-agent orchestration. ModusMax’s own dual path — cloud APIs and local models — reflects a broader industry truth: choice and control beat one-size-fits-all.

5. Ship, learn, narrow

Applied AI products improve when scope is narrowed after contact with reality. He has argued publicly that hiding ideas on a laptop helps no one — while still protecting what truly must stay confidential.

Positioning without mythology

Market conversations increasingly value practitioners who can sit with design, ML, and delivery — the niche of corporate design AI and AI product engineering. That demand is real. What this profile will not do is invent league tables, fake client logos, or vanity metrics. Credibility in Applied AI is earned the old way: clear thinking, working software, and honest communication.

Teams hiring or partnering for Applied AI / AI product design in South Africa or remotely can judge the work: the product thesis, the UX of agency, the operational seriousness.

Who this is for

Product and design leaders who need AI features that survive contact with users.
Engineering managers who want design-literate AI builders, not slideware.
Founders exploring agent platforms and automation without drowning non-technical users in CLI friction.
South African teams who want global-quality AI product craft with local context.

Connect

For conversations on Applied AI, agentic workflows, or AI-native product design:

LinkedIn: Delon Swartz
ModusMax: modusmax.online
TheJobPlug: thejobplug.co.za

Applied AI in South Africa does not need more mystique. It needs designers and engineers who ship carefully — and keep learning in public.

About

Delon Swartz is an AI Design Engineer / AI Product Engineer & Automation Specialist based in Durban, South Africa. He focuses on Applied AI, Applied ML, and AI product design; founded ModusMax AI (desktop AI agents); and builds end-to-end digital products through Delon Labs UX. Connect on LinkedIn.