Alejandra Osuna
Frontend Engineer
production-ready React/Next.js
Building pixel-perfect, accessible interfaces with a product mindset.
From design systems to shipping features that matter.
Selected Work
A few projects that show how I think about UI, production quality, and shipping.
Local-first RAG system that makes AI conversations queryable and reusable. Hybrid semantic search (text + vector), a ranking arbiter, multilingual embeddings, AES-256-GCM encryption, and an MCP server connected to Claude and Gemini CLI.
Note: Some details are intentionally general due to NDA.
Fullstack recycling rewards app with 27 real Bogotá recycling centers, Clerk auth middleware, Leaflet map integration, Supabase RLS policies, and auth-gated data access. Built to encourage sustainable habits in the city.
Note: Some details are intentionally general due to NDA.
Chrome extension that tackles algorithmic echo-chamber bias by surfacing diverse perspectives. Built with Plasmo framework. Live at breakthebubble.dev. Tutorial published on Programming Witch YouTube channel.
Note: Some details are intentionally general due to NDA.
Professional website for Venezuelan psychologist Mariale Hernández. Includes a fully functional admin panel with Supabase auth, CRUD management for workshops (talleres) with image upload, and a public-facing landing page. Deployed on Vercel.
Note: Some details are intentionally general due to NDA.
Experience
Fullstack Engineer
Spybee · Remote · B2B SaaS · Construction Monitoring
Built the largest feature of the quarter from scratch and architected an internal AI chatbot — working across the full stack from React frontend to Hono/Node.js backend contracts.
- Architected the Daily Logs module from scratch across 80+ files — dynamic log types, configurable fields, PDF/Excel export, role-based permissions, and a custom form builder — largest feature shipped that quarter.
- Built an internal AI chatbot integrating OpenAI embeddings, MongoDB Atlas vector search, and Claude API into the Next.js frontend, enabling semantic Q&A over operational data.
- Refactored monolithic Potree 3D point cloud viewer into clean custom hooks, reducing component coupling and enabling safe parallel work by 2 engineers.
- Detected an inconsistent user data pattern across the codebase, built a shared UserName/UserDisplay component, and rolled it out in 4 locations with zero regression.
- Built i18n audit pre-commit script and Driver.js onboarding tour; authored PR review guidelines adopted by the team.
Frontend Engineer
Emote Care · Remote · HealthTech · UK-based · Platform closed Jan 2026
Owned end-to-end feature development on a mental health platform serving therapists and patients across multiple regions — from requirement clarification through production release.
- Built and maintained a reusable UI component library with Tailwind CSS and React, reducing time-to-implement new screens by ~30%.
- Migrated 14+ landing pages from Svelte to Next.js and platform content to Strapi CMS, handling component architecture and routing independently.
- Delivered complex multi-step user flows (onboarding, therapist-patient matching, verification) with strong accessibility and UX focus on a two-sided marketplace.
- Improved reliability through Playwright and Testing Library coverage; maintained CI/CD pipelines via Azure DevOps and Docker across a distributed UK team.
Some implementation details are intentionally general due to NDA and privacy constraints.
Skills
React / Next
- React 18+
- Next.js 14 (App Router)
- TypeScript
- Server Components
- React Hooks
State & Data
- Zustand
- TanStack Query
- Zod
- Context API
- SWR
UI Engineering
- Tailwind CSS
- shadcn/ui
- Ant Design v5
- Framer Motion
- Responsive Design
Backend & Data
- Node.js
- Hono
- PostgreSQL
- Supabase (RLS, Edge Functions)
- MongoDB Atlas
- pgvector / RAG
AI & Integrations
- Claude API (Anthropic)
- OpenAI API + embeddings
- MCP Servers
- Streaming UI
- Claude Code · Cursor · Gemini CLI
Testing & Tooling
- Jest
- Playwright
- Testing Library
- GitHub Actions
- Docker
- i18n (next-intl)
Impact
Production Features
Shipped user-facing UI with reliability and performance in mind — from 0 to production, end-to-end.
AI-Powered Systems
Built RAG pipelines, vector search, and LLM integrations (Claude, OpenAI) into production products.
Reusable Systems
Created component libraries and shared patterns that scale across features and reduce bugs.
Cross-functional Work
Worked closely with design, product, and backend across US and UK distributed teams.