Junior Frontend & Full-Stack Developer · Creative Technologist
I build React, TypeScript, Python and SQL products with accessible multilingual UX, data honesty and practical automation.
💼 LinkedIn · 📄 CV EN · CV ES · CV HE · ✉️ Email · 🎧 Nova Music Lab live · א Ivrit Sheli live
Open to: Junior Frontend & Full-Stack roles · Born in: San Cristóbal, Venezuela · Based in: Beersheba, Israel · Languages: ES · EN · HE
I’m Kevin Cusnir, born in San Cristóbal, Venezuela and now based in Beersheba, Israel. I build React, TypeScript, Python and SQL products with accessible multilingual UX, data honesty and practical automation.
Open to junior frontend, full-stack and creative-technology opportunities. Best fit: a junior team with mentorship, real users, thoughtful review and room for structured creativity.
| What I can prove publicly | Strongest evidence |
|---|---|
| React and TypeScript product work | Nova Music Lab, Ivrit Sheli and Christopher Rodríguez Portfolio |
| Full-stack, Python and data persistence | Ivrit Sheli v2.4.0 with FastAPI/PostgreSQL; NovaFit v4.2.0 with Python/SQLite |
| Data, accessibility and multilingual UX | Honest source-aware analytics; EN/ES/HE, RTL, keyboard and reduced-motion work |
| Delivery discipline | 213-test full-stack proof, Docker, CI, live deployments, bundle budgets and release checks |
Review path: 30 seconds: Read the snapshot and role positioning · 2 minutes: Open Nova Music Lab and Ivrit Sheli · 5 minutes: Inspect tests, CI, security architecture and recent releases · 15 minutes: Run Ivrit Sheli or NovaFit, or explore the live music museum.
Status: Deployed v1.5.0 local-first flagship · Stack: React 19 · TypeScript · Vite · Dexie/IndexedDB · Recharts · Vitest · Playwright
Problem → solution: Listening exports are fragmented, hard to compare and easy to misrepresent. A local-first music museum that turns listening exports into a Living Artist Atlas, source-aware genre evidence, emotional stories and a portable archive while keeping private imports in the browser.
Evidence: Deployed on 2026-08-01 at commit 5c1d57048b52, with 711 passing unit/integration tests, 18 passing Playwright journeys, privacy audits, bundle budgets and a verified GitHub Pages smoke test · Role signal: Frontend engineering · data visualization · privacy-aware product thinking
Open Nova Music Lab live demo · Nova Music Lab source
Released product — v1.5.0, deployed 2026-08-01: the preview was captured from the final private candidate before publication; the live manifest, exact commit and CI verify the working deployment.
🎧 Open the Nova Music Lab 1.5.0 product tour
Release evidence: candidate media captured 2026-08-01 and included in deployed commit 5c1d57048b52; the static frame above replaces animation when reduced motion is preferred.
Five import families become one deduplicated, source-aware listening history. Missing fields remain visible as gaps, and raw exports stay in the browser.
Explore the live music museum · Inspect the Nova Music Lab source
Status: Live v2.4.0 dual-mode full-stack product · Stack: React 19 · TypeScript · FastAPI · Python · PostgreSQL 17 · SQLite · Alembic · Docker · Railway
Problem → solution: Hebrew learners need focused practice, progress and multilingual guidance without surrendering private study data or depending on a fragile cloud service. A private-first trilingual Hebrew-learning product with a guided contest tour, 48 reviewed visual concepts, local SQLite, authenticated PostgreSQL, native RTL and accessible motion.
Evidence: Verified Railway production and PostgreSQL readiness for release implementation 03bf84b9268f, with 151 backend + 62 frontend = 213 passing tests; identity-only Google sign-in, session refresh, onboarding persistence, logout and signed-out reload are verified, while a successful live GitHub session and re-login after logout remain unverified · Role signal: Full-stack architecture · backend security · PostgreSQL ownership · integration testing · multilingual RTL UX
Open Ivrit Sheli live demo · Ivrit Sheli source
א Open the archived Ivrit Sheli 2.2.0 product tour and verified 2.4.0 full-stack proof
Public-data boundary: The sanitized live captures use Kevin's isolated account and show his already-public name/avatar plus default recommendations (12 words, one oral practice and 10 minutes); they contain no private study history, credential, token, provider result or production database record.
Visual evidence boundary: These 2.2.0 screens are interaction history, not visual proof of the live 2.4.0 interface.
Archived Ivrit Sheli 2.2.0 interface: These frames document the deployment captured on 2026-07-18 and do not claim visual proof of the newer release implementation 03bf84b9268f.
Verified v2.4.0 evidence: 151 backend + 62 frontend = 213 passing tests · Railway production · PostgreSQL 17 ready · live/ready health checks true · release implementation commit 03bf84b9268f · tenant RLS · Alembic · non-root Docker · redacted structured JSON logs.
OAuth boundary: Identity-only Google sign-in, onboarding/session persistence across reload, logout and signed-out persistence after reload are verified in production. Re-login after logout, a live GitHub account session, live OpenAI or Google Workspace connector calls, two-real-user isolation and backup restoration remain unverified; Google sign-in grants no Gmail, Drive or Calendar scope.
Publication boundary: the verified deployment, Git tag and GitHub Release now agree on v2.4.0.
Inspect the Ivrit Sheli source · Review the test contract · Inspect the reviewed upstream project snapshot · Open the archived Hebrew RTL frame · Open verified live deployment
Status: Active v4.2.0 local-first desktop product · Stack: Python · Tkinter · SQLite · Requests · Faker · Matplotlib
Problem → solution: Daily wellness data should remain portable, understandable and private. A local-first Windows wellness intelligence studio with isolated profiles, English, Spanish and Hebrew RTL, efficient motion, explainable analytics, complete verified backups and one-click self-repair.
Evidence: 124 discovered automated tests, 12 themes, EN/ES/HE RTL UX, verified complete backups, 58 public visual assets, one-click verification and strict release audit · Role signal: Python application architecture · desktop UX · SQLite migrations · i18n/RTL · analytics · release engineering
Open NovaFit live demo · NovaFit source
💙 Open the NovaFit 4.2.0 product tour and verified evidence
Public-data boundary: The tour and static fallback use clearly labeled seeded demo profiles and deterministic synthetic records; they contain no real wellness history, runtime database, secret or private export.
NovaFit 4.2.0 in motion: A concise recruiter-facing desktop tour demonstrates navigation, motivation design, explainable coaching, profile isolation and multilingual RTL behavior; the static evidence below documents the analytics and twelve-theme system.
Manifest-backed product summary: A local-first Windows wellness intelligence studio with isolated profiles, English, Spanish and Hebrew RTL, efficient motion, explainable analytics, complete verified backups and one-click self-repair.
Verified v4.2.0 evidence: 124 discovered automated tests · 12 themes · 58 public visual assets · EN/ES/HE with Hebrew RTL · complete verified backups · installable static showcase · one-click Windows release.
The Training Atlas is profile-independent; the theme contact sheet uses seeded demonstration records without publishing a profile name.
12 curated themes: Midnight Neon · Aurora Borealis · Negev Sunrise · Ocean Depth · Forest Focus · Rose Quartz · Cloud Day · Solar Paper · High Contrast · Royal Sapphire · Cyber Lime · Sunset Arcade.
Open the NovaFit live showcase · Download the verified v4.2.0 release · Inspect the NovaFit source · Verified project manifest
🌍 Open NovaFit profiles, EN/ES/HE, coach and safe delivery
Public-data boundary: this system map contains no profile names, dates or wellness metrics.
Each profile owns isolated records, goals, language, theme and activity preferences. English and Spanish use LTR; Hebrew moves the shell to RTL. Suggestions expose data confidence and reasons while avoiding medical claims.
The checker repairs a local .venv, validates Matplotlib and Asia/Jerusalem, runs 124 discovered automated tests, preserves an existing user database in workspace mode, and uses VERIFY_ALL.bat plus python tools/package_audit.py --strict-distribution for reproducible verification and clean downloadable releases.
Status: Client/collaboration case study · Stack: React · TypeScript · Vite · Tailwind CSS · Framer Motion
Problem → solution: A real educator needed a maintainable bilingual professional presence. An accessible React and TypeScript portfolio with structured content, verification states, persistent themes and GitHub Pages delivery.
Evidence: EN/ES content architecture, keyboard UX, reduced motion, SEO and automated deployment · Role signal: Client communication · accessible frontend · maintainable content architecture
Open Christopher Rodríguez Portfolio live demo · Christopher Rodríguez Portfolio source
Status: Audited six-week learning archive · Stack: Python · JavaScript · TypeScript · Node.js · SQL
Problem → solution: Course exercises need context, progression and reproducible quality checks. A structured archive from Python and OOP through JavaScript, DOM, async workflows, TypeScript, Node and databases.
Evidence: Six curriculum weeks, 14 discoverable test files, CI, generated audits and transparently documented remaining quality gates · Role signal: Learning progression · problem solving · Git/PR discipline
Fullstack2026 source
Evidence counts come from the featured public projects and their documented quality pipelines.
| Evidence across featured work | Count |
|---|---|
| Featured products/case studies | 5 |
| React or TypeScript projects | 4 |
| Python evidence areas | 3 |
| Multilingual/accessibility projects (including RTL) | 4 |
| Automated quality pipelines | 4 |
Languages: TypeScript · JavaScript · Python · SQL
Frontend: React · HTML5 · CSS3 · Tailwind CSS · Vite · Accessibility · i18n / RTL
Backend & data: Node.js · Express fundamentals · REST APIs · SQLite · PostgreSQL · FastAPI · Alembic · JSON / CSV
Testing & quality: Vitest · Testing Library · unittest · pytest · integration testing · structured logging · GitHub Actions · linting · data audits · bundle budgets
Workflow: Git · GitHub · Docker · Conventional Commits · README-first documentation · AI-assisted review with explicit verification
Life route: San Cristóbal, Venezuela → Beersheba, Israel
Country boundaries and deterministic tiny-state markers represent 195 sovereign states while framing a personal journey from Venezuelan roots to an Israeli home, supported by Spanish, English and Hebrew communication.
| Language | Level | Product value |
|---|---|---|
| Spanish | Native | Clear communication and Spanish localization |
| English | Advanced professional | Technical documentation and international collaboration |
| Hebrew | Working local proficiency | Israeli workplace communication and RTL product testing |
Next depth: Production monitoring, service-level objectives and alert design · PostgreSQL backup, restore and migration drills · Authorization, OAuth and browser-level end-to-end testing · Free-tier deployment, capacity and cost operations
🧠 Engineering approach and transferable strengths
- Persistent debugging and practical troubleshooting
- Accessible multilingual product design, including RTL behavior
- Data transformation, validation and honest handling of uncertainty
- Clear technical documentation and visual product storytelling
- Local-first privacy decisions and proportional security claims
- Responsible early-adopter testing of preview and beta software in non-critical contexts, with reproducible feedback aimed at improving future releases
- Make the user-facing claim proportional to the evidence.
- Keep inputs validated and failure states actionable.
- Treat accessibility, privacy and documentation as implementation work.
- Prefer small reversible changes, focused commits and reproducible checks.
- Use AI to accelerate review and exploration without outsourcing understanding.
🎓 Education and structured learning
Full-Stack Development — Developers Institute, Israel
2025–2026
Coverage: Python, OOP, SQL, JavaScript, React, Redux, Node.js, Express, JWT, TypeScript, Git/GitHub and deployment.
Current status: Completing and converting remaining coursework into tested, deployable portfolio evidence.
The public learning archive remains separate from production-ready projects so recruiters can distinguish progression from finished product evidence.
🧙♂️ Kevin and Lirioth — professional structure and creative signature
Kevin is the professional engineering identity: implementation, debugging, documentation, responsibility and collaboration.
Lirioth Teltanion is the creative signature: music technology, visual language, narrative systems, experimental interfaces and generative art.
Together, these identities make technically sound products easier to understand, remember and enjoy.
- 💼 LinkedIn — Kevin Cusnir
- 📄 English CV · Currículum en español · קורות חיים בעברית
- 🐙 GitHub — Lirioth Teltanion
- ✉️ Email Kevin
Code with purpose. Design with personality. Data with honesty. 💙