JV_
EN ES

JESUS VAZQUEZ / JEALVH

Full Stack Developer building reliable web products and practical AI integrations.

I turn ambiguous product needs into durable systems: clear interfaces, explicit rules, observable workflows, and human review where it matters.

See selected workStart a conversation

WORKING PRINCIPLES

SELECTED WORK

Work that can be explained, not just shown.

Some repositories are private, in beta, or part of a team effort. These notes focus on responsibilities, system boundaries, and verified behavior without exposing confidential implementation details.

Status is intentional: shipped behavior and in-progress integration are kept distinct.

01 / BETA PROTOTYPE

Tailored Parts

Assisted quotation prototype for auto-parts shops.

Transforms ambiguous parts requests into structured searches, preserves conversational context, and routes uncertain cases to human review.

  • The main branch verifies an admin-only WhatsApp Cloud API flow, PostgreSQL persistence, and durable message recovery.
  • Telegram, a normalized catalog, QuoteFlowService, and an optional-LLM parser are in integration and not presented as shipped.
  • AI is deliberately bounded to natural-language interpretation; deterministic rules and the catalog remain the source of truth.
PythonFastAPIPostgreSQLWhatsApp Cloud APITelegramOptional LLM

02 / TEAM PLATFORM

AMG RFID Tools

Industrial tool traceability platform using RFID.

Contributed to a platform connecting a React/TypeScript frontend, Go backend, PostgreSQL, and an edge gateway on Raspberry Pi.

  • Verified platform capabilities include gateway synchronization of tools and tags, entry and exit resolution, location and status updates, and movement records.
  • Exit destination resolution falls back from kanban_zone to the tool destination when needed.
  • The legacy gateway code includes TCP RFID reading and checksum validation. This is team platform work, not a claim of sole authorship.
ReactTypeScriptGoPostgreSQLRaspberry PiRFID

CAPABILITIES

Full-stack foundations, chosen for operational clarity.

Web products

  • TypeScript and modern frontend systems
  • Python/FastAPI and Go services
  • Relational data with PostgreSQL
  • Static, accessible, performance-minded interfaces

Integration work

  • Messaging APIs and durable event handling
  • Edge gateways and device-facing workflows
  • Structured search from unstructured requests
  • Explicit fallbacks and observable state changes

AI-ASSISTED ENGINEERING WORKFLOW

AI assists the process. Fundamentals govern the result.

I use Gentle AI, Engram, and OpenCode to support planning, implementation, and continuity of context. The workflow stays grounded in system boundaries, verification, and human review rather than treating AI output as authority.

This is an engineering workflow, not an AI product or an AI Engineer claim.

  1. 01Frame the problem and constraints
  2. 02Implement in small, verifiable units
  3. 03Check behavior, states, and assumptions
  4. 04Keep a human accountable for the decision

OTHER PROJECTS

More signals from the workbench.

sparring-pro

Conversion-focused boxing gym landing page.

Astro / TypeScriptPrivate

toyoda-venado

Door-profile verification system v4 for a Mitsubishi FX3G-60M PLC.

Industrial automation

CONTACT

Have a product problem worth structuring?

I am available for conversations about reliable web products, integrations, and the systems around them.

Email JesusView GitHub