Three products, a small team: how Lemus Digital operates

The question we get most

Every time someone lands on lemusdigital.com and sees three business lines —our own SaaS products, AI training and implementation, and professional streaming infrastructure operations— the question is always the same:

"Isn't that too many things for a small team?"

It's a good question. The short answer: yes, it would be too much if they were three generic businesses. But they're three lines that share exactly what matters: the same technical stack, the same way of thinking about product, the same hands in production.

This post explains why that structure is deliberate, how it holds up day to day without becoming chaos, and which decisions we made early that we wouldn't touch.

What "three products" actually means

Before defending the structure, it's worth clarifying what's really under the umbrella. "Three lines" can sound like three teams, three P&Ls, three meeting rooms — that's not it.

Line 1 — Our own SaaS products

Assilek is our first product in the air. SaaS for professional photographer proofing galleries. It lives under the LLC, has its own domain (assilek.com), its own infrastructure on GCP (Next.js + Spring Boot + an image-service that processes images with watermarks via Pub/Sub), its own blog, its own support channel. It's a serious product, with subscription, Stripe checkout, adaptive multi-currency pricing, and 24/7 operation.

Building and operating Assilek is real work. Not a decorative side project.

Line 2 — Aiuda Labs (generative AI)

This line is the most interesting structurally, because Aiuda Labs is a DBA —a registered fictitious name— of Lemus Digital LLC. Legally it operates under the LLC; a client signing with "Aiuda Labs" is signing with Lemus Digital LLC. It's led by my brother Noel Moreno Lemus, PhD in Data Science, from Panamá, with full operational autonomy. He works with founders and teams in LATAM adopting applied AI — workshops, coaching, implementation.

The key to the setup: shared legal structure, autonomous operational direction. Aiuda has a clean legal home under the LLC, and Noel keeps full authority over how the line is built and operated. Below I explain why this line entered under the LLC and the streaming one hasn't yet.

Line 3 — Streaming infrastructure

The know-how for satellite capture, Flussonic processing, Stalker portals, geographic failover — all of that lives in my head and in active projects I maintain outside the formal LLC umbrella. It's the least formalized line today, on purpose: I'm evaluating how to formalize it well before putting it under the LLC with clients and public pricing.

What goes under the LLC, and what doesn't (yet)

The logical next question: if the same two people run all three lines, do they all belong to a single entity?

The answer is nuanced. Today two of the three lines live under Lemus Digital LLC —Assilek directly and Aiuda Labs as a registered DBA— and the third, streaming, stays outside on purpose. The reasons:

1. Each line has a different legal risk profile

A multi-tenant SaaS that stores private photos and an AI consultancy share a manageable risk profile —software and professional services— so they coexist fine under the same legal structure. Operating streaming infrastructure with third-party transmission rights is something else: a legally distinct surface. That's why streaming doesn't enter under the LLC yet — we want to isolate that risk until we formalize the service properly. Important to note: a DBA doesn't create liability separation. Aiuda Labs shares the legal exposure of Lemus Digital. That's why we only put under the same roof what has compatible risk.

2. Aiuda Labs has its own identity, even sharing the legal structure

Noel built Aiuda Labs from Panamá with his own sweat equity, his own network, his own clients. Registering it as a DBA of Lemus Digital LLC gives it a clean legal home and brand clarity without erasing any of that: Noel keeps full operational authority over the line. We share the legal structure; the day-to-day direction is his. It's the difference between "subsidiary controlled from Florida" and "a line with a clear operational owner under a common legal umbrella".

3. The streaming line isn't ready to have public clients under the LLC

The technical expertise is there. What's missing is deciding how to formalize the service (hourly pricing? per project? monthly retainer? what SLA level for clients?). Putting it under the LLC without those decisions made would mean offering a half-baked service, which is exactly what we don't want to do.

What we do share (and why that's the strength)

Beyond who operates each line, what unites them? The short answer: what matters.

The same technical stack

Noel and I program in the same languages (Java, Python, Scala). We use the same tools (Claude Code, Cursor, MCP, Next.js, Postgres, vector DBs when applicable). When one of us solves a technical problem in our line, the learning transfers to the other with zero friction. When Assilek needs a RAG pattern for search, I know Noel has been through it on some Aiuda project. When Aiuda needs an opinion on how to architect a service that will scale, he knows I lived that problem building Assilek.

It's the difference between two people working in silos and two people sharing a library of solutions. And that library wouldn't be available if each of us were an island.

The same brand voice

The three channels (Lemus Digital, Assilek, Aiuda Labs) talk differently in detail, but share the same principle: technical, honest, no fluff, no buzzwords. If we say it, we built it. If we didn't, we say that too.

That shared voice is what lets each blog feel familiar to the reader coming from another. It's not centralized marketing — it's editorial coincidence because the philosophy is the same.

The same definition of "what we are not"

None of the three lines is a generalist agency. None sells "integrated digital solutions". None presents itself as an expert in everything. Each is small and specific:

• Assilek does one thing (proofing galleries for photographers) and does it well.
• Aiuda Labs works with applied generative AI — training, coaching, implementation. None of that "digital transformation" talk.
• Streaming is technical infrastructure for specific clients (legitimate operators, churches with live broadcasts, internal corporate video). Not "audiovisual services".

That shared discipline of staying specific is what prevents three small lines from turning into a mediocre mid-sized agency.

How we coordinate day to day without overhead

The practical question: if structures are separate and products are independent, how do we coordinate without burning hours on meetings?

No recurring meetings

No daily stand-ups, no weeklies, no planning. When one of us needs the other's technical opinion, we write — text, voice note, screenshot. If it needs a conversation, we schedule that one conversation. The rest of the time each of us works on our line.

This sounds obvious but it's the difference with most small teams that recreate the overhead of large companies "to be professional". Meetings exist to manage coordinated uncertainty. When two people have clarity about who leads what, most meetings disappear.

Technical decisions: whoever executes decides

If Noel is implementing something in Aiuda Labs and has to choose between two architectures, he decides. If I'm deciding how to handle images in Assilek, I decide. We ask for opinions when there's time and the decision is cheaply reversible, and we consult when it's expensively irreversible. But opinion isn't vote — whoever executes is the one who knows the details the other doesn't see.

Brand: explicit conventions, not centralized rules

There's a brand context master document (the file lemus-digital-contexto.md in this same repo, which the observant reader will find) where the decisions that don't get re-litigated live: voice, positioning, what we're not, technical attributions, separation between what's LD and what belongs to the child brands. That document is the source of truth. When in doubt, consult the document. When the document is wrong, update it explicitly.

It works because it fits in a single scroll. If it needed a whole wiki, we'd already have too much complexity for the team size.

What we wouldn't do again

It would be dishonest to make it look like we got to this structure by divine inspiration. There were mistakes. Three worth naming:

1. Starting to communicate before the structure was clear

Early on we communicated "Lemus Digital" without the legal structure being resolved behind it. That generated ambiguity: it wasn't clear which entity an Aiuda Labs client was contracting with. We fixed it by formalizing — Aiuda Labs is today a registered DBA of Lemus Digital LLC, so the client knows exactly who they're signing with. Structure has to come before brand, not after.

2. Underestimating the cost of running streaming in parallel

Keeping an IPTV/OTT headend alive isn't "monitor when the alarm goes off". It's constant background work — TBS card drivers that change with every kernel update, operator subscriptions that renew on weird dates, EPGs that break from timezone changes, alerts at 3 AM. You can't treat it as an occasional task. The deliberate decision to keep it separate from the LLC was, in retrospect, the only thing that allows Assilek to have the focus it needs.

3. Wanting to "professionalize" too early

At some point we considered bringing in management tooling, quarterly OKRs, team metrics. None of that applies to two people. Energy must go to product and clients, not to processes. Management tools serve when the bottleneck is coordinating ten people. With two, they're noise. Professionalism isn't bureaucracy.

For whom this is replicable

This structure isn't for everyone. It works if three conditions are met:

1. There's a founder with clear technical authority in each line. If the line doesn't have an owner who can build and operate without supervision, the structure doesn't scale — it breaks at the first human bottleneck.
2. There's philosophical alignment, not just goal alignment. Sharing voice, definition of quality, criteria for "what we don't do" is what keeps three lines as an ecosystem and not as a random collection. If the people have different philosophies, the umbrella brand feels forced to the reader.
3. There's discipline to say no. Each line will get opportunities outside its scope. The temptation to "take the project because it pays well even if it's a bit out of scope" is what turns boutiques into mediocre agencies. Saying no is the most underrated operational skill.

Closing

Lemus Digital is a structural experiment: three lines that look like an agency but operate like a small lab. The way they stay united where it gives leverage (stack, voice, cross-learning, and a common legal structure where the risk is compatible) and autonomous where it matters (each line with its operational owner and its own criterion) is what makes it sustainable.

If you're thinking about building something similar — an umbrella with several small, well-done products instead of a generic mid-sized product — it's worth taking the time to design the structure before communicating it. What's cheap to do at the start is expensive to undo later.

If you want to compare notes on your own structure, or explore how we could collaborate between one of our lines and yours, let's talk.