Skip to content
ESEN
Portrait of Timothy Kuno

Available for remote roles and freelance projects

Full-stack developer · Applied AI

Hi, I’m Timothy Kuno.

I design and build full-stack systems, automation, and applied AI for problems that exist beyond the screen.

Download CVLet's talk
Cochabamba · Bolivia · RemoteScroll to rebuild

MANIFESTO / 00

Software with good handwriting: systems that are clear, maintainable, and work beyond the screen.

Selected work

CAF / One dataset, four perspectives

CAF / One dataset, four perspectivesFour roles query one shared dataset. RLS returns four, three, two or one packet according to permissions. In the field, GPS and QR verify a visit; the mobile queue retains visits offline and synchronizes when online returns. Conceptual simulation with no patient data.
Queue ready to synchronize
CAF admin dashboard with test data: module menu for operations, inventory, and finance, plus command-center notices.

This is where it all starts.

01 / 05Web · Mobile · Backend · Security

CAF — Integrated system

An institutional digital ecosystem for managing cases, sessions, inventory, and field work through one architecture.

Read the case
Context
The organization needed to replace fragmented workflows while handling clinical, legal, and financial information with granular access.
System
I architected a Next.js platform, a PostgreSQL/Supabase model with RLS, and an offline-first React Native app for GPS- and QR-verified visits.
Outcome
The system runs in production as a multi-module platform with role-based traceability and a complete delivery lifecycle.
My role
Full-stack engineer responsible for architecture, data, web, mobile, and delivery
  • Next.js
  • TypeScript
  • Supabase
  • PostgreSQL
  • React Native
  • Expo
  • RLS

Emotional AI / Signal and context

Emotional AI / Signal and contextFace, biometrics and eye tracking converge in acquisition and a multimodal model to estimate affective state. A RAG branch retrieves laboratory data to answer a question with evidence. Dashed lines identify context sources.
Architecture simulation

Every session, recorded and traceable.

02 / 05Multimodal ML · RAG · Real time

Emotional AI and RAG assistant

An ecosystem combining acquisition and inference for facial, biometric, and eye-tracking signals with conversational queries over experimental data.

Read the case
Context
The lab relied on a legacy biometric stack and disconnected analyses, limiting both real-time research and data exploration.
System
I led the tuning of a multimodal model and its acquisition pipeline, built a live facial-inference installation, and developed a RAG assistant that retrieves relevant context before answering in natural language.
Outcome
The foundation connects research, live demonstrations, and queries grounded in evidence available in the lab's data.
My role
AI direction, technical leadership, and RAG pipeline implementation
  • Python
  • Machine Learning
  • DeepFace
  • OpenCV
  • WebSockets
  • Eye Tracking
  • RAG
  • LLM
  • Structured Output
  • Data Pipelines

MQTT + gRPC / Dispatch and return

MQTT + gRPC / Dispatch and returnThe UI sends a task to the master. The master publishes to upb/workers/tasks through MQTT and distributes to three workers, one busy. Workers return results via the gRPC SendResult callback to the master, which updates the UI. Dispatch and return follow separate paths.
Architecture simulation

A query enters the system.

03 / 05Distributed systems · Real-time communication

Distributed MQTT + gRPC system

A master/worker architecture that dispatches tasks over MQTT and returns results through gRPC callbacks to a web interface.

Read the case
Context
Concurrent processing required dispatch, execution, and result delivery to stay separate without coupling workers to the UI.
System
I defined message and service contracts to coordinate workers, callbacks, and state visible from a web application.
Outcome
The project demonstrates asynchronous communication, clear service boundaries, and coordination across distributed components.
My role
Distributed architecture design and implementation
  • TypeScript
  • MQTT
  • gRPC
  • Web UI
  • Master/Worker
View repository

Jarvis / One voice, two routes

Jarvis / One voice, two routesVisitor voice becomes text through speech recognition. RAG retrieves mall data; the LLM composes a response and the avatar speaks it. With cloud disabled, the transcript takes a local response route back to the avatar voice, bypassing the remote LLM.
Responding via cloud

Hi, I'm Paseito.

04 / 05Conversational assistant · Voice · Web application

Paseito, Paseo Aranjuez assistant

A responsive web application with an avatar and voice for asking about Paseo Aranjuez's catalog, hours, promotions, and events.

Read the case
Context
The challenge called for bringing catalog, hours, promotions, and event queries into a conversational experience.
System
I connected the avatar, voice, catalog, hours, promotions, events, WhatsApp, and analytics to the same service; the portable version uses browser text and speech.
Outcome
The public demo is served over HTTPS; the tower keeps local GPU voice, with localhost as an in-person fallback.
My role
Responsive application, backend, and shared conversational service
  • FastAPI
  • SQLite
  • OpenAI Responses
  • Browser speech
  • Whisper
  • Kokoro
  • Docker
Open demo

Nexo / From deliveries to decisions

Nexo / From deliveries to decisionsTXT deliveries pass through an importer that checks duplicates and overlaps before writing to PostgreSQL. The dashboard reads metrics and displays charts. Agent tools read metrics and modify the board. Demonstration data is fictitious.
Fictitious demonstration data

Your numbers, clear at first glance.

05 / 05Commercial dashboard · Data · AI assistant

Nexo Dashboard TXT

A commercial dashboard that processes TXT deliveries into PostgreSQL and combines interactive charts with an AI assistant to query metrics, explain results, and edit the board through conversation.

Read the case
Context
Commercial delivery data arriving in TXT files needed to become queryable from a dashboard.
System
The importer processes TXT files from a folder into PostgreSQL. The assistant queries metrics, explains results, and edits the board through conversation; the frontend and backend work independently of Odoo.
Outcome
Functional demo with fictitious data.
My role
Commercial dashboard, delivery importer, and AI assistant
  • TXT
  • PostgreSQL
  • Interactive charts
  • AI assistant

More projects

Experience

2024 — PresentCochabamba, Bolivia

Director of Artificial Intelligence

Neuroscience & AI Laboratory · UPB

I lead the evolution of the lab's AI infrastructure and connect biometric research with usable products.

  • Multimodal emotion-prediction model
  • RAG analytical assistant
  • Real-time acquisition pipeline
Jan. 2026 — Jul. 2026Remoto

Data & Automation Analyst · Internship

Bamboo Tec

I designed Python ETL pipelines and executive Power BI dashboards for operational data.

  • Python and Pandas ETL automation
  • Corporate KPI monitoring
Dec. 2025 — Aug. 2026Sucre · Remote

Full-stack Developer · Freelance

Family Support Center · CAF

I independently built a multi-module institutional platform and its companion mobile application.

  • Next.js and Microsoft Graph
  • Supabase with RLS
  • Offline-first React Native
Sep. 2025 — Dec. 2025Cochabamba, Bolivia

Automation Engineer · Internship

Trilogy Software · VIVA

I automated technical support and enterprise integrations for infrastructure ticket lifecycles.

  • n8n orchestration
  • REST and webhook integrations

Capabilities

Applied AI

  • LLM fine-tuning
  • RAG
  • Structured output
  • Multimodal machine learning

Frontend & web

  • React
  • Next.js
  • TypeScript
  • Tailwind CSS
  • Responsive design
  • SEO

Backend & data

  • Node.js
  • Python
  • C#
  • PostgreSQL
  • Supabase
  • ETL

Integration & automation

  • REST
  • Webhooks
  • WebSockets
  • n8n
  • Puppeteer

Cloud & DevOps

  • AWS
  • Vercel
  • Docker
  • Git
  • CI/CD

Mobile

  • React Native
  • Expo
  • Kotlin
  • MVVM
  • Offline-first

Education & achievements

2022 — 2027

Computer Systems Engineering

Universidad Privada Boliviana · UPB

Coursework in machine learning, data engineering, and distributed systems, with active technical leadership in the Neuroscience & AI Laboratory.

IEEE Xtreme 2023

5th place nationally

ICPC Bolivia

Regional competitor

AWS

Cloud Practitioner and AI Practitioner

IEEE

Co-organizer of IEEE Xtreme 2023 and the 2024 IEEE Student Branch Meeting

Let's build something that works in the real world.

I'm open to remote roles and freelance projects where product, data, and engineering need to work together.

  • Employment opportunities
  • Freelance projects
juchpaul20@gmail.com