santiago
amaya
product engineer · ai-native systems · full-stack execution
building ai-driven and operational software for real-world business environments. erp, pos, automation, internal tools — combining engineering, delivery, and pragmatic ai to ship systems that work.
experience
mar 2023 – present
capybara solutions
co-founder & product engineer
- built and operated software products and custom systems across commerce, operations, and automation use cases
- designed and deployed production systems using fastapi, postgresql, docker, and modern frontend frameworks
- integrated llm workflows and ai-assisted tooling for content generation, automation, and faster internal execution
- led end-to-end product lifecycle: problem framing, architecture, implementation, deployment, and iteration
- worked directly with business needs under ambiguity, translating rough requirements into functioning systems
may 2022 – feb 2023
texpar
full-stack engineer
- improved landing page seo, contributing to a 60% increase in organic traffic
- integrated a chatbot using the openai api, improving user experience and conversion rates
- reduced form completion time by 35%, boosting conversions by 25%
- built reusable ui components with next.js, sass, and ant design to accelerate feature delivery
- improved page load speed from 4.2s to 1.8s through structured data and meta tag optimization
selected systems
restaurant pos platform
01- full point-of-sale system with offline-capable workflows, backend sync, and thermal printer integration
- surrounding platform with backend services, hosted menus, and a web dashboard for operators
- real-world constraints across local operation, syncing, deployment, and multi-tenant concerns
custom erp platform
02- custom erp from scratch — schema design, backend structure, and workflow modeling for business operations
- normalized sql data models and implementation patterns for operational reliability
- technical direction, execution, and broader product/system decisions
capy cards
03- ai-powered study tool generating quizzes and flashcards from pdfs
- practical document transformation and educational workflows
- content generation pipelines and subscription-oriented product direction
memocap
04- event photo-sharing product — submit and access media through a web experience
- architecture across frontend, backend, object storage/cdn, auth, and deployment
- practical integration of uploads, hosted delivery, and user access flows
scancat enrichment
05- enrichment workflows extracting business information from public sources
- automated restaurant/business onboarding with structured data collection
- combined scraping/enrichment logic with operational usability
internal tools & automation
06- monitoring, deployment utilities, syncing workflows, and small internal systems
- ai coding tools and agent-style workflows for day-to-day implementation and iteration
capabilities
python · fastapi · postgresql · docker · typescript · javascript · sql · react · next.js · svelte · tailwind · rest apis · auth flows · system design
llm integration · prompt workflows · agent-style systems · automation pipelines · document transformation · structured data workflows
linux · cloudflare · caddy · gitlab ci/cd · vps deployment · monitoring · reverse proxies · containerized environments
sqlalchemy · alembic · electron · scraping/enrichment workflows · thermal printer integration
build philosophy
use ai as a practical development multiplier — prototyping, implementation, debugging, workflow automation
validate and harden ai-assisted output for production reliability instead of treating generated code as correct
combine deterministic software systems with llm layers when they improve usability or decision-making
prioritize shipping speed, real constraints, and iteration over unnecessary process
make architecture and tooling decisions based on leverage, maintainability, and delivery pressure
education
in progress · 2028
b.s. mechatronics engineering
unlz
software, electronics, automation, control systems, and applied engineering foundations
i focus on building systems that solve real operational problems — turning incomplete requirements into working software that can actually be deployed, used, and iterated on.