M2Labs / personal lab
Facundo Matias Aquino · personal project portfolio

Software for real life, built with AI.

I'm Facundo Matias Aquino, an Applied AI & Automation Builder. I build small, useful systems for problems I actually have — then keep improving them through use.

See selected work LinkedIn ↗
EU citizen · open to opportunities in Ticino, the rest of Switzerland, and Italy
M2Labs / selected workFORJA WODS · HEALTHY POCKETS
Forja WODs home screen showing the selected day and workout controls
Forja WODs
Healthy Pockets dashboard showing fictional Demo Portfolio values and asset allocation
Healthy Pockets
Forja WODs · Healthy PocketsTwo mobile products, shown at equal scale.
01 / selected work

Three projects. Real problems.

The projects below are personal applications. They are not presented as startups or commercial products — they are working systems built around daily use.

01 · mobile application

Forja WODs

A personalized workout generator, training log and per-profile RM percentage calculator. Start with your own custom profile and optionally add others. Choose a focus, level and RX/Scaled mode, generate a WOD with optional AI or built-in fallback, run the workout timer and record the result.

AI-assisted generation Fallback logic Training state Mobile product

Design decision: the app stays useful without an API key. A user who does not know what an API key is can still generate a WOD; users who want more personalization can connect an OpenAI or Anthropic model.

Open Forja WODs web app ↗ Download Android APK ↓
Forja WODs home screen with day selection, training mode and workout statusForja WODs · home and day view
Forja WODs one-repetition-maximum percentage calculator screen
01 / RM calculator
Forja WODs warm-up and workout sections with bottom navigation
02 / workout sections
Watch the original screen recording
02 · local-first application

Healthy Pockets

A local investment-tracking application that brings positions, movements, imports and projections into one place. It is connected to Hermes for reviewed agent proposals, while keeping the user in control.

Local-first CSV / Excel import Agent review Scenario modeling

Privacy scope: Screenshots use fictional demo values only. Healthy Pockets is a personal tracking and analysis tool; it does not connect to brokers or execute real trades.

Open Healthy Pockets web app ↗ Download Android APK ↗
Healthy Pockets dashboard showing a fictional USD 21,600 Demo Portfolio with asset allocationDashboard · fictional USD 21,600
Healthy Pockets input method with local investment and transaction options
01 / local-first input options
Healthy Pockets profile and privacy settings with synthetic demo profile
02 / profile & privacy settings
Watch the original screen recording
03 · playable web demo

First 100 Days

A demo of the Argentine Presidential Simulator: an educational pixel-art game where you take office on 10 December 2023 and have 100 days to make 7 decisions on the exchange rate, public spending, social aid, regional investment, a union dispute and food prices. Every option shows its benefit and its cost: there are no right answers, only priorities.

Deterministic engine No server or accounts Offline PWA SVG pixel art

Scope: this is a playable demo, not the full game. Indicators start from data inspired by late-2023 figures, but outcomes come from a simplified game model: they are not forecasts or real policy. The game does not use AI at runtime.

Play the demo in your browser ↗
Pixel-art presidential office with the seated president and three ministers; a side panel shows the Economy minister's opinion on the first decisionOffice · synthetic “Demo” save
Decision screen with three exchange-rate options, each with its benefit and cost, and the button to sign the chosen option
01 / decision with benefit and cost
Final report of the first 100 days listing the indicators that improved and worsened for President Demo
02 / day-100 report
02 / the shift
“The business logic is all going to these agents.”

— Satya Nadella, Microsoft CEO · BG2 podcast · December 2024

My reading: this is not the end of software. It is a move away from one-size-fits-all applications toward tools shaped around a person, a workflow or a specific context — made practical by AI.

Original episode ↗Transcript excerpt ↗
03 / how I build

Useful first.
Fancy later.

My laboratory background shapes how I approach software: define the inputs, check the output, document the process and make failure visible.

  1. 01
    Start with a real problemThe product begins with a workflow I actually need, not with a tool looking for a use.
  2. 02
    Keep the happy path simpleA useful system should still make sense to someone who does not know the technical vocabulary behind it.
  3. 03
    Design for the failure caseWhen an AI response is slow, invalid or unavailable, the user should know what happened and what to do next.
  4. 04
    Do not trust output too quicklyTesting, boundaries and documentation matter as much as the first impressive demo.
04 / about

From the lab bench to applied AI.

I started my career in clinical laboratory work, where reliable results depended on process: protocols, controls, documentation and repeat checks. Today I apply that mindset to small AI-enabled products and automation workflows.

I am currently pursuing a Master's in Artificial Intelligence at BIG School, while continuing to build and use my own tools. I am interested in practical AI systems, agent workflows, API integrations and implementation work where careful thinking matters.

FocusApplied AI · agents · workflow automation
BackgroundClinical laboratory · QA/QC · SOPs
LocationAdelaide now · Europe next
DirectionTicino · rest of Switzerland · Italy
05 / contact

Let the work speak first.

Connect on LinkedIn ↗