Reverse engineering
Independent study of software behavior, internal structures and real-time data. Private research and source code stay private.
Tomas BarzelattoReverse engineering · Programming · Product
Understand
the system.
Build the
experience.
Independent reverse engineering, programming and interactive product work. I explore how software behaves and turn complex ideas into usable interfaces.
Explore selected work
Preparing interactive study
Drag to orbit. Select a layer to explore.
A system worth exploring.
Three layers connect systems thinking, interface design and product development. This is a real-time visual study, not a hardware product.
Project preview
Concept diagram. No live integrations.
A spatial way to explore projects, manual time logs and report drafts.
Read the case studyFrom understanding
to implementation.
Three connected parts of the way I work.
Programming
Web interfaces, visualizations, audio experiments and practical tools, with an emphasis on understandable behavior and useful controls.
Product development
Defining requirements, shaping the experience, evaluating results and iterating with AI-assisted engineering.
A city for my
AI workflow.
Each building represents a role, from research to review. Explore an example of how I organize work with dots and GPT agents.
Interactive render · Simulated activityPreparing city
Choose Explore city, then drag to orbit or Shift-drag to pan. Zoom in to watch the traffic. Finish exploring to restore page gestures.
From a question to a next step.
Select a building to explore its role in the example workflow.
The city, camera and moving cars are implemented in the browser. Agent roles and handoffs are an illustrative workflow, not live agent telemetry or private project activity.
Selected work.
Concrete ideas.
A selection of product ideas and practical interfaces. Each case separates what exists from what still needs to be built.
Ñami
Meal planning for Chile · Web app
A week of meals, planned around everyday life.
A Spanish-language meal-planning web app for Chile, connecting weekly menus, recipes and organized shopping lists in a mobile-first experience.
Built with AI assistance, with my focus on product direction, requirements and iterative UX feedback.
Explore ÑamiDecisions & boundaries
The product decision
Bring meal ideas, weekly planning and the shopping list into one flow, with attention to portions, preferences, pantry staples and phone-friendly navigation.
Current boundary
The linked project is a web app. Its native iPhone app is marked coming soon. Fridge scanning and checkout are not presented here as verified working integrations; prices are estimates.
Spatial work dashboard
Interactive workspace · Prototype
What if organizing work felt more like navigating a place?
An isometric city brings project templates, manual time entries and report drafts into one interface. The canvas supports orbit, zoom and a reset view, while conventional lists keep the same work accessible.
Decisions & boundaries
The product decision
Use the city as another way into the work, while retaining lists and direct controls. Keep time entries manual rather than presenting guesses as recorded work.
Current boundary
The visible projects and activity are examples. The prototype does not show live agents or connected workplace activity.
Contribution
Product direction, visual reference selection, requirements and iterative feedback; implementation developed with AI assistance.
Learning playground
Language practice · Prototype
Start with curiosity. Make language practice part of the conversation.
A learning concept combines topic selection, adjustable correction styles and session recaps with an original melody and a tap-timing activity.
Choose a subjectMath
Correction styleCoach as we go
“I am agree.”
Try “I agree.”
Illustrative excerpt from a deterministic correction rule, not a live AI response.
Decisions & boundaries
The product decision
Let the learner choose the subject and how often to receive corrections. Keep conversation, learning and music in one coherent product rather than separate utilities.
Current boundary
Conversations are scripted and correction rules are limited. No AI model, microphone, speech recognition, pronunciation scoring or vocal-pitch assessment is connected. The rhythm activity measures taps only.
Contribution
Original product concept, learning experience requirements and iteration priorities; AI-assisted interface, implementation and test preparation.
Cat-finding interface
Spatial guidance · Simulation
Explore the finding experience before connecting the hardware.
A direction-and-distance prototype uses a virtual target and manual heading to explore nearby, behind-you and further-away states. An optional camera background lets the interface be considered in context.
Decisions & boundaries
The product decision
Separate interaction design from uncertain hardware access. Make simulated values unmistakable and preserve usable controls when the camera is unavailable.
Current boundary
There is no live AirTag, GPS, Bluetooth or Ultra Wideband connection. The marker is not spatially anchored. This is an interface simulation, not a tracker or a safety device.
Contribution
Problem definition, desired finding experience and hardware questions; AI-assisted simulation, interface and test planning.
Beyond prototypes
Venue planning
Calendar-driven business modeling, scenario comparisons and role-based views. The underlying project uses TypeScript, Vite and Supabase. Private business data and source are omitted here.
Client websites & operations
Company and venue websites, bilingual presentation and event administration workflows. Product requirements and repeated UX feedback center on making everyday tasks easier.
A clear idea.
A testable next step.
I work with AI throughout the build process. My contribution includes product ideas, requirements, design direction, feedback and iteration. Implementation and test preparation are AI-assisted.
- Define the useful thing
- Turn an idea into concrete interactions, constraints and examples that can be evaluated.
- Build, inspect, refine
- Work with AI on implementation, then use feedback and visible behavior to guide the next iteration.
- Keep the boundaries clear
- Distinguish sample data, simulations and future capabilities from what a prototype actually does today.
The background
behind the work.
Computer scientist with experience in business management and inventory operations. Based in Greater Santiago, Chile.
View LinkedIn profileExperience
Software Developer & Web Developer
GOZO AI
General Manager
Mastercajas S.A.S
Inventory Specialist
Jorge Barzelatto & Cia Ltda
Education & recognition
Lindenwood University
University studies
Dean’s Honor Roll
Recognition for Spring 2019 academic achievement.
Community
Delta Tau Delta
Initiated active member
Tools behind the work.
Languages and tools reflected in my projects and public professional profile.
- Programming languages
- Python · C++ · JavaScript · Java · TypeScript · AngelScript
- Web foundations
- HTML · CSS
- Interactive systems
- Canvas · Web Audio · UI state & browser events
- Build & delivery
- Node.js · Vite · GitHub · Vercel
- Data & persistence
- Supabase · Local and cloud state
These projects involve AI-assisted engineering. Repository ownership is not presented as proof that every line was written independently.
Good software starts
with a good question.
Explore the work, the decisions behind it and the technologies used to bring it to life.
Back to the top