AI agents that hold up in production.
I am Marco Fura. I take AI and ML systems from a stakeholder conversation to production: scoping loosely defined problems with enterprise customers, building the agents and pipelines that solve them, and changing the platform when the field says it is wrong.
- Salesforce
- PagerDuty
- Pontificia Universidad Católica de Chile
- Universidad Nacional Mayor de San Marcos
Experience
From founding a company to enterprise AI.
Nov 2026 – Present
Salesforce
Lead Forward Deployed Engineer
Working closely with enterprise customers to take AI agent projects to production end to end, and leading the internal improvements that raise how the team performs.
Sep 2025 – Sep 2026
PagerDuty
Forward Deployed Engineer
Embedded with enterprise customers and internal teams, turning operational needs into production infrastructure and applied AI.
Feb 2025 – Aug 2025
PagerDuty
Innovation Software Engineer
Led innovation projects end to end: re-architecting a customer-facing platform into microservices and micro-frontends to shorten release cycles, and building data lake infrastructure on AWS.
Dec 2022 – Jan 2025
Katari
Software Engineer
Built AI systems that reached production, from retrieval-augmented generation over enterprise records to fine-tuned sentiment models. Shipped applications for international B2B and SaaS clients.
Dec 2021 – Mar 2024
Prysmo Solutions
Founder & CEO, then Software Engineer
Founded the company and ran technology projects end to end: scoping, budgeting, stakeholder alignment and delivery. Mentored engineers on Clean Architecture, design patterns and TDD.
Aug 2021 – Nov 2021
Senergy
Software Engineer
Implemented a forecasting model for electricity consumption and spend, built for the energy sector.
Education and research
MSc
Artificial Intelligence
Pontificia Universidad Católica de Chile · 2024–2025
Thesis: EVA, a knowledge-management framework for LLM systems on Knowledge Graphs, advanced RAG and the Model Context Protocol.
BSc
Software Engineering
Universidad Nacional Mayor de San Marcos · 2017–2022
Thesis research on multi-speaker voice cloning: neural network strategies and models for personalized speech synthesis.
Publication
Neural Network Strategies and Models for Voice Cloning in a Multi-speaker Mode: An Overview
Springer · ICIT 2022
Book chapter, presented at ICIT 2022 and published by Springer.
Expertise
- AI and ML
- AI agents · Agentic architectures · MCP · RAG · Knowledge Graphs · Evaluation and benchmarking · Forecasting · Clustering · Fine-tuning · PyTorch · TensorFlow
- Backend and architecture
- Python · Java · TypeScript · FastAPI · Spring Boot · Microservices · Micro-frontends · Event-driven systems
- Cloud and DevOps
- AWS · Terraform · Kubernetes · Docker · CircleCI · GitHub Actions · Backstage · Atlantis
- Data
- PostgreSQL · MongoDB · Qdrant · pgvector · ETL and ELT pipelines · Data lakes
- Observability
- OpenTelemetry · OpAMP · SNMP