A
Clarity
See the whole system before touching a single part. Most failures are drawn into the map long before they ship.
Independent systems studio / Est. 2019
GA4MMA
Systems that create leverage.
We design the product, intelligence and capital systems behind ambitious operations, and we build them to keep standing.
46° 03′ 20″ N / 14° 30′ 30″ EScroll to descendGamma / 07
Gamma builds systems that create leverage.
Effort adds.
Systems multiply.
Every ambitious operation hits the same wall: more output demands more people, more hours, more heroics. Leverage is the way through it.
A system costs more at the start. Then it crosses the line and keeps paying out in users, in time and in capital. We design for that crossing.
Guardian doctrine
Clarity / Resolve / StewardshipB
Resolve
Take the hard part first. That is where the leverage hides, and where most teams stop looking.
C
Stewardship
Build it to hold: through scale, through change, through the migration nobody planned for.
We draw the whole map.
Each of these turns a tangle of people, data and value into something you can actually operate. Watch what moves through them.
A food habit, ~10k strong.
An owned consumer product built around one daily question: what should I eat today? Structured menus, regional discovery and a loop that brings people back.
Search that understands intent.
A RAG architecture that doubles as a search engine, steered by our two-layer semantic control protocol so answers stay grounded in live data.
Many institutions. One core.
Multi-tenant learning: each institution keeps its own identity, roles and isolated data while the platform evolves as one system.
Follow the money all the way through.
An order architecture where every handoff belongs to one settlement system. Nothing gets lost between customer, merchant and ledger.
Proof that survives the upgrade.
Signed, verifiable attestations of knowledge, versioned so tomorrow's migration never breaks yesterday's signature.
From inbox noise to operation.
Inbound messages become structured records, queued and delivered on schedule. Raw signals, turned into controlled operations.
Bring us the map you can't draw yet.
Written before it's built.
Every system starts as a document that has to survive hard questions. Here is one, opened: the AI retrieval and vector search system behind our food platform.
Doc / G4-SPEC-AI-01AI processing & vector search systemRev / 2026 Specified
Select a section
01.0 / Purpose
Turn structured food data into natural-language discovery.
People don't search for restaurants. They search for a mood, a budget, a place, a time. The system translates that intent into grounded answers drawn from menus that change every day.
- “burgers under €15”
- “lunch for vegetarians”
- “open after 22:00”
- “somewhere with kids”
- Scope
- ~200 restaurants, daily-changing menus
- Primary question
- What should I eat, here, now?
- Design rule
- AI improves the journey quietly, never as a gimmick
From the archive
- 01AI processing & vector search
- 02Multi-tenant LMS platform
- 03LMS security review
- 04Proof-of-knowledge attestation
- 05Token withdrawals & deposits
- 06Webhook mailing architecture
- 07White-label course platform
- 08In-browser course IDE
- 09Subscription & ordering system
- 10Mobile portfolio app
- 11Game design document
- 12Capital strategy overview
+ more in the vault. Specifics stay with the people they belong to.
9:41●●●
jedelbom
What are you in the mood for today?
burgers under €15, lunch for vegetarians, open after 22:00, somewhere with kids, spicy Asian food
- Today's lunch3 courses / €8.90 / 400 mOpen now
- Wood-fired pizzaTerrace / until 23:001.1 km
- Vegetarian bowlUnder €12 / quietDaily
Illustrative interface
Jedelbom / owned product
What should I eat today?
Jedelbom is our own food platform: structured daily menus from local restaurants across Slovenia's regions, with AI that quietly improves every search. There's no AI tab, just better answers.
~10kUsers, approaching
~200Restaurants in AI scope
12Regions mapped
DailyMenus refreshed
“AI should quietly improve everything.”
Jedelbom design brief
We wrote it down in 2019.
Our Global Market Report was written in the aftermath of the 2018 sell-off: nine markets, a one-year and a five-year horizon, and theses committed to paper. Here is how they aged.
20.6×
ETH, from $237 at writing to its November 2021 high.
9
Markets analysed: equities, gold, oil, USD, EUR, CNY, BTC, ETH and uranium.
5
Calls tracked against what the market did next.
1+5
Year horizons: the next twelve months, and the five years after.
| Market | June 2019 | The call | Later reached | Multiple |
|---|---|---|---|---|
| Ethereum | $237price at writing | “Smart contracts give ETH real utility. It may prove the better investment than Bitcoin.” | $4,878Nov 2021 | 20.6× |
| Bitcoin | $8,700price at writing | “A digital store of value with mathematically provable scarcity.” | $100,000Dec 2024 | 11.5× |
| Uranium | $17/lbcycle low cited | “Grossly undervalued. Energy demand will make it matter again.” | $100/lbJan 2024 | 5.9× |
| Gold | $1,300price at writing | “Higher lows, ascending triangle: a generally bullish structure.” | $3,000Mar 2025 | 2.3× |
| S&P 500 | 2,750index at writing | “The 2018 sell-off was fear, not fundamentals. No structural decline around the corner.” | 6,000Nov 2024 | 2.2× |
Report by Sanel Okić, published 3 June 2019. “Later reached” shows public market milestones after publication, not portfolio returns. Uranium is measured from the cycle low cited in the report. Past performance does not predict future results.
Grow the stack. Never sell the principal.
Turn ETH into yield-bearing collateral, borrow conservatively against it, and dollar-cost average that liquidity back into ETH. Stay invested, stay liquid, and compound with the trend instead of guessing it.
- 01
Stake
ETH becomes yield-bearing collateral. The base grows on its own.
- 02
Borrow
Stablecoin liquidity against it, at a conservative target LTV.
- 03
Accumulate
Dollar-cost average every week. A system, not a prediction.
- 04
Trend
As ETH rises, collateral grows and LTV falls on its own.
- 05
Restore
Borrow only back to target. Risk stays constant, never escalates.
- 06
Unwind
In rallies, sell DCA ETH, repay the loan, keep the surplus.
- ↻
Repeat
Small, repeated, safe actions. Compounding does the rest.
Net carry
- Staking yield
- +3.2%
- Borrow cost
- −3.4%
- Net
- ≈ 0
Borrowing becomes a liquidity tool, not a cost center.
Stress test / illustrativeTarget LTV 25%
-35%
Loan-to-value38.5%
Drawdown absorbed. Principal untouched, and each DCA buy gets cheaper.
Stack over a trend cycle
Principal, never sold Accumulated
What the system refuses
- Timing the market
- Selling the principal
- Borrowing past target
- Big-dip heroics
Working thesis from our internal ETH executive strategy overview. Rates and ratios are illustrative assumptions at the time of writing. This is not financial advice or an offer of any financial service. Leverage carries liquidation risk.
Digital, Future, Ideas, Reality, Hope, Think, Create, Gamma, G4mma
A step in the right direction.
Recovered from our unfinished 2019 homepage
Four chapters, one name.
The step became a system.
We started in 2019 building websites, frontends and Electron apps for clients in the US. Each chapter widened the scope, from interfaces to architectures to operations. The fourth carries the number.
- 2019
Gamma Digital
Websites, frontends and Electron apps for US clients. Hands-on delivery.
- Catalyst
Gamma Catalyst
Architecture beyond frontend code. Our first multi-tenant web3 education platform.
- GCS
Catalyst Solutions
Holistic delivery of complex operations at rapid iteration and quality.
- g4Now
g4mma.digital
An owned product near 10k users, an intricate RAG engine, capital research. Leverage.
Bring us the
hard part.
We take on a small number of systems each year. If it is complex, consequential and still unfinished, we should talk.
Join our Discord