Open to remote / nearshore opportunities

Strategy · Architecture · Delivery

Hugo Antonio Segoviano MartínezAI Solutions Architect

LLM agents, RAG, automation, AI Product Engineering, and technical leadership.

I turn ambiguous business problems into production-ready AI, software, and integration solutions—connecting discovery, architecture, and hands-on execution.

Explore AI solutions Download résumé Contact me on WhatsApp

Portrait of Hugo Segoviano, AI Solutions Architect

AI delivery system

From business problem to operable solution

DiscoveryUse case, risks, and value signal

ArchitectureData, models, integrations, and controls

ProductFull-stack implementation and evaluation

ProductionDeployment, observability, and iteration

Working modeArchitecture + hands-on

6+years of professional experience

15+countries using customs software

≈40%reduction in page-load time

99.8%uptime maintained

Selected AI solutions

From use case to a system that can operate.

Five cases showing how I connect business context, architecture decisions, implementation, and evidence without overstating the maturity of each solution.

Production01

Novonet Customer Support AI Agent

Conversational customer-support agent for an Ecuadorian internet provider.

Problem / context: Turn support workflows into a useful conversational experience for real customers.
Solution / architecture: Customer-support flows translated into an operable conversation and response solution.
My responsibility: Solution design and deployment of the customer-support agent.
Status / evidence: Used in real customer interactions. The approved sources do not document volume or resolution rate.

Operational flow

Support workflows

Conversational agent

Real interaction

Functional implementation02

SegoMart Telegram Sales Agent

Sales agent connected to Telegram and real operating data.

Problem / context: Query products, pricing, inventory, and customers in real time from a sales conversation.
Solution / architecture: Tool calling orchestrated with n8n and OpenAI, Telegram as the interface, and Firestore as the production-data source.
My responsibility: Agent architecture, security controls, trade-off analysis, and design of a 10-case evaluation set.
Status / evidence: Functional implementation; the set was designed, without claiming that all 10 cases passed or that the system is fully production-hardened.

Real-time lookup

Telegram

n8n + OpenAI

Firestore data

Public repository · Functional03

RAGAgent

End-to-end document RAG assistant for querying PDFs with grounded answers.

Problem / context: Make PDF content queryable through semantic retrieval and conversational context.
Solution / architecture: Layered TypeScript/Express backend; PDF ingestion and chunking, Mistral embeddings and generation, Pinecone, semantic retrieval, in-process session memory, and document-level attribution.
My responsibility: End-to-end design and implementation of ingestion, retrieval, generation, and chat experience.
Status / evidence: Functional project and public repository. It is not presented as having an active public deployment.

View repository

RAG pipeline

PDF ingestion

Mistral embeddings

Pinecone + answer

Public repository · Functional04

FinBot

Contextual financial-guidance agent for Mexico with persistent memory and tool use.

Problem / context: Offer conversational guidance that uses a financial profile, history, and current-information search.
Solution / architecture: FastAPI and LangChain with OpenAI, Firestore memory, a Tavily tool, SSE streaming, a React client, Docker, and Railway configuration.
My responsibility: Full-stack implementation of the agent, profiles, memory, tool use, streaming, and web experience.
Status / evidence: Functional project and public repository. Railway configuration is included, but the previous deployment is not active.

View repository

Contextual agent

Profile + memory

Agent + Tavily

SSE + React

Product with customer05

AI-powered ERP for SMBs — HiveDev

Multi-module ERP connecting POS, sales, inventory, and automation for Latin American SMBs.

Problem / context: Unify fragmented operating processes and turn business needs into a useful, sellable product.
Solution / architecture: Multi-module full-stack product with AI and automation components, designed from discovery through cloud operation.
My responsibility: End-to-end ownership: discovery, product definition, architecture, development, cloud, and client iteration.
Status / evidence: First paying customer within 60 days of the MVP.

Product cycle

Discovery

Multi-module product

Client iteration

Professional experience

Architecture, product, and delivery in real environments.

A career spanning founder-led products, international B2B platforms, and mission-critical customs and logistics software.

Founder & AI Solutions Architect / Product Engineer
HiveDev
Jan 2020 — PresentMexico · Remote
I design software, AI, and automation solutions for Latin American SMBs with full ownership of the product lifecycle.
- Launched a multi-module AI-powered ERP and secured the first paying customer within 60 days of the MVP.
- Designed agent solutions for support, sales, financial guidance, and RAG-based document retrieval.
- Lead discovery, architecture, implementation, cloud deployment, and direct client iteration.
- React
- Next.js
- React Native
- TypeScript
- Node.js
- Python
- FastAPI
- PostgreSQL
- Firebase
- AWS
- Railway
Tech Lead / AI Developer
Proceti IT · Simultrayd project
Dec 2025 — Jul 2026Arizona, USA · Remote
Led technical delivery for an international B2B platform focused on global trade, logistics, automation, and applied AI.
- Coordinated a distributed team of 4–6 developers and converted executive needs into executable roadmaps.
- Designed integrations and capabilities involving voice agents, automated email, conversational flows, and dashboards.
- Orchestrated commercial and operational workflows through APIs, webhooks, and business rules.
- Make
- Knack
- Stripe
- Slack
- Bland.ai
- SendGrid
- Apollo.io
- Hunter.io
- TrestleIQ
Full Stack Developer
CustomsCity
May 2022 — Aug 2024Ontario, Canada · Remote
Developed mission-critical international customs and logistics software used across 15+ countries.
- Delivered 5+ production features for thousands of daily users and more than USD 10M in transaction volume.
- Reduced page-load times by approximately 40% while contributing to 99.8% uptime.
- Collaborated across three countries and helped mentor developers.
- React
- Node.js
- PostgreSQL
- MongoDB
- Redis
- AWS
Full Stack Developer
Thincode
Jan 2020 — Jun 2020Mexico
Built internal systems that improved project visibility and tracking.
- Delivered two management systems and reduced tracking overhead by approximately 50%.
- Ruby on Rails
- PostgreSQL
Software Development Intern
Universidad de Guanajuato
Jan 2019 — Dec 2019Mexico
Developed a system to manage academic research projects and reduce administrative overhead.
- Built the solution with C#, .NET, SQL Server, and Azure.
- C#
- .NET
- SQL Server
- Azure

How I build solutions

From ambiguity to executable architecture.

I combine product thinking, hands-on engineering, and leadership to make technical decisions grounded in business context.

Solution and product architecture

Discovery, use-case definition, MVP scope, risks, data, APIs, and integrations.

LLM applications, agents, and tool calling

Agentic flows with tools, context, controls, evaluation, and explicit reliability decisions.

RAG, embeddings, and semantic search

Ingestion, chunking, vectorization, retrieval, grounding, memory, and source attribution.

Automation, APIs, and integrations

End-to-end workflows using webhooks, OAuth, external services, and business rules.

Full-stack, data, and cloud

Web and mobile products, backends, persistence, deployment, and cloud operation.

Technical leadership and stakeholders

Roadmaps, trade-offs, documentation, mentoring, and direct communication with executive and operational teams.

Consolidated stack

Technology selected for the problem—not the trend.

AI

Automation

Backend

Frontend

Data

Cloud and DevOps

Continuous learning

Business, artificial intelligence, and engineering.

A technical foundation complemented by active study in business strategy and AI solution architecture.

Education

Master of Business Administration (MBA)Tecnológico de Monterrey
2025 — In progress
Master's in Artificial IntelligenceUniversidad Da Vinci
2025 — In progress
B.S. in Computer Systems EngineeringUniversidad de Guanajuato
2020
Computer TechnicianCBTis 65
2015

Certifications

Certified AI/LLM Solution ArchitectBSG Institute
2026
Scrum Foundation
2024
Generative AIUdemy
2023
Advanced React & Next.jsUdemy
2022

About me

Architecture with judgment; execution with context.

I do my best work where technology, product, and business need to speak the same language.

I combine solution architecture, hands-on execution, and product thinking. I can move an initiative from an ambiguous discovery conversation to a deployed, evaluated solution ready to iterate.

I have collaborated with teams and clients across Mexico, the United States, and Canada, working directly with executive leadership, operations, and distributed developers. That experience helps me balance speed, reliability, maintainability, and business value.

Understand before buildingThe use case, data, and operating reality define the architecture.
Stay close to the codeI design solutions I can also implement, evaluate, and improve.
Communicate the trade-offsTechnical decisions should be clear to both business and engineering.

Contact

Let's turn a complex need into an executable plan.

I am open to AI Solutions Architect, AI Product Engineer, and Tech Lead roles, as well as remote or nearshore collaborations. Message me directly on WhatsApp.

Message me on WhatsApp

WhatsApp+52 462 266 5621

LinkedInlinkedin.com/in/hugosegoviano

GitHubgithub.com/hugohasm

Websitehugosegoviano.com

LocationMexico · Remote / Nearshore