Available for full-time remote

Amara Torretti

Full-Stack & Agentic AI Engineer

I architect and build local-first AI systems, full-stack products, open-source tools, and interactive experiences that give people more privacy, control, and creative agency.

PythonTypeScriptReactFastAPIRustNode.jsSQLiteSupabase
View engineering work View résumé Contact me
1Merged upstream
OpenClaw contribution
54Indexed projects
across the portfolio
7+Years across
software and IT
4+Live browser
demos and games
100%Remote-first
by design

Featured engineering work

Real systems. Real implementation. Clear boundaries.

View GitHub
Open sourceMerged

OpenClaw PR #22142

Cross-surface UI integrity fix

Traced AI-only metadata leaking into user-visible conversation history across webchat, TUI, and Swift surfaces. Implemented a fast stripping parser, regression tests, and a clean resubmission that merged upstream.

TypeScriptSwiftTestingUI systems
View merged pull request
Full-stack productActive

HumanPalette

Verified-human art marketplace

A React Native marketplace combining swipe discovery, realtime chat, artist verification, Stripe Connect onboarding, escrow-style payments, Supabase RLS, and six Edge Functions.

React NativeSupabaseStripeTypeScript
Explore product architecture
Agent systemOperational core

MIST / clawd

Hybrid local + cloud AI runtime

A mobile-to-gateway companion architecture spanning React Native, TypeScript, Python WebSocket services, Ollama, cloud model routing, durable memory, host awareness, and consent-gated tools.

PythonReact NativeWebSocketOllama
Read the case study

Selected systems

Depth across security, browsers, agents, and proof.

These projects earned placement through implementation evidence, not just concept scale.

01

Ghostline

Operational modules

CLI-first defensive security suite with seven independently installable modules for scanning, password breach checks, DNS leak testing, public-data recon, web audits, account recovery checks, and bookmark auditing.

PythonCLIPytestSecurity
7 modulesOpen repo
02

AetherBrowser

Publishing-ready core

A Tauri v2 desktop shell with React and TypeScript, dual Studio and Arcade modes, local persistence, project creation, a publishing checklist, and an embedded agent workspace.

RustTauriReactTypeScript
Phase 1–2 completeOpen repo
03

AetherProof

Challenge prototype

Voice-and-vision proof workflow for human artists, designed around Gemini Live, React, Node.js, Firestore, Cloud Run, and Solana devnet certificates.

Gemini LiveReactCloud RunSolana
Architecture scaffoldedOpen repo
04

AetherTwin Fairy OS

Local-first MVP

Embodied agent-design environment with Genesis profiles, visible strengths and shadow risks, a skill registry, council decisions, permission boundaries, handoff briefs, and a browser-persisted memory tree.

ReactAgentsLocalStoragePermissions
MVP shippedOpen repo

Play & interactive demos

Systems you can touch, not just read about.

No account required. Open a world, test a loop, or explore an interface.

Playable browser game

Pet Grave

Cozy-haunted spirit pets, persistent bonds, day/night combat, procedural audio, and mobile controls.

3D living world

Aether Garden

Explore a glowing world, bond with spirits, and watch autonomous builders raise shrines.

Agent simulation

QI Games

Observe a hybrid agent society whose decisions combine utility logic with quantum-inspired scoring.

Web audio instrument

Boop Beat Board

A 16-pad beat board and step sequencer with a Web Audio engine, mobile shell, and haptics bridge.

Project atlas

The full repo constellation, organized for signal.

Loading index… The strongest work leads; prototypes and archives remain discoverable without taking over the page.

About

A systems thinker who refuses to separate engineering from imagination.

I bring enterprise IT experience, software engineering training, full-stack product work, open-source contribution experience, and years of building across AI, web, mobile, security, and games.

My best work happens where architecture and interface meet: turning complicated agent behavior, local data boundaries, tool permissions, and persistent state into something a person can actually understand and use.

Read my résumé Explore GitHub

How I work

  1. 01
    Truth before theater

    I label prototypes honestly, show what is operational, and avoid inflating roadmap ideas into shipped capabilities.

  2. 02
    Agency by design

    Permission boundaries, local-first options, exportability, and human approval are product features, not legal footnotes.

  3. 03
    Receipts over claims

    Tests, merged pull requests, live demos, architecture diagrams, and reproducible setup instructions carry the story.

  4. 04
    Memorable, then usable

    A product can have soul without sacrificing clarity, accessibility, or operational discipline.

Open a channel

I’m looking for a remote team where ambitious AI products become real software.

Best fit: full-stack AI, agentic product engineering, developer tooling, local-first systems, creative technology, or interactive product teams.

Email Amara View résumé GitHub