Diego Moreno: AI/ML Engineer and Data Scientist
As a solutions architect, I design, build and deploy data and AI systems that solve real problems and run in production in the cloud, improving decisions, optimizing operations and creating business value. I work from defining the problem to putting the product into operation.
Projects, consulting, contractor engagements and job opportunities.
Areas of expertise
- Forecasting and optimization
- Agents and integrations
- Data platforms
- AWS
Solutions applied to real problems
A selection of my work in forecasting, artificial intelligence and operations. Each project shows the problem addressed and my contribution.
Forecasting for consumer goods in Latin America
Improving demand forecasting in Latin America with PepsiCo's global data science team.
- Demand history
- Models and evaluation
- Demand planning
- My contribution
- Work on forecasting models and their evaluation to support demand planning.
- Stack
- Azure Databricks
- Microsoft Fabric
- Azure Data Factory
- Python
As a contractor through Insight Global, I worked with PepsiCo's global data science team to improve forecasting for Latin America, alongside professionals from Europe, India and the region, in an enterprise environment of Azure Databricks, Microsoft Fabric and Azure Data Factory.
- Problem
- Improve the demand forecast that feeds planning in Latin America.
- Team
- Global data science team with professionals from Europe, India and the region.
- Engagement
- Contractor through Insight Global.
Backend for an operational query assistant
Backend for a chatbot that queries merchants, terminals and tickets, with RAG for frequent questions.
- User question
- Orchestration and RAG
- Operational data and FAQ
- My contribution
- Backend development and integration of the queries and the knowledge retrieval component.
- Stack
- AWS Lambda
- Amazon Lex V2
- Amazon Bedrock
- RAG
- Python
I built most of the backend of a chatbot for a financial services organization. It lets users query merchant, terminal, user and ticket-status information, and uses RAG to answer frequently asked questions.
- Problem
- Give conversational access to operational information and frequently asked questions.
- Solution
- An orchestration layer that routes each query to the operational integration or to knowledge retrieval.
- Client
- Financial services organization, anonymized.
Optimizing commercial and operational decisions
Promotion optimization that connects data science with business constraints.
- Commercial data
- Optimization model
- Promotion decision
- My contribution
- Contributed to developing the optimization solution.
- Stack
- Optimization
- Python
- SQL
I took part in a promotion optimization project for a commerce and delivery platform. The work connected data science with commercial decisions and business constraints.
- Problem
- Choose promotions while respecting commercial goals and operational constraints.
- Approach
- Data science connected to the business's rules and constraints.
Other work in data, AI and operations. Consulting clients remain anonymous.
More projects
Steel industryIn development
IoT analytics and logistics for a steelmaker
Batch pipeline for plant sensor data with validation and silver and gold layers for querying and dashboards, plus a logistics stream covering deliveries, scheduling and fleet.
AWS Step Functions · AWS Glue · Lambda · Athena · QuickSight
AgribusinessMVP
Weekly production forecast for an agricultural exporter
Multi-model forecast by horizon (short, medium and long term) with rolling backtesting, ETL, processing and an API in the cloud.
SageMaker Processing · AWS Glue · Athena · Terraform · Power BI
Manufacturing
Industrial data with a conversational assistant
Industrial sensor data flowing into a data lake and dashboards, with an assistant that answers questions in natural language.
AWS IoT SiteWise · AWS Glue · Athena · Amazon Lex · Amazon Bedrock
Demand planning
End-to-end automated forecasting engine
Replacing Excel and manual notebooks with a staged pipeline: extraction, cleaning, features, models, ensemble, reviews and automatic upload to the ERP.
Python · SQL Server · Airflow · Microservices · RPA
Commerce and delivery
Product-level demand forecasting
Product-level forecasts for four cities, connected to purchasing and inventory planning.
Python · SQL · AWS · Git
Floriculture
Order allocation and customer segmentation
Data warehouse with information from companies in several countries, customer clustering and linear programming to allocate orders.
Linear programming · Clustering · QlikView · SQL
Commerce and delivery
Operational space
Space allocation optimization for dark stores and warehouses.
Optimization · Python
Retail
Retail productivity and scheduling
Productivity, inventory and cashier scheduling models, plus support for in-store space management.
Power BI · QlikView · SQL
Personal product
Forecastor
Web app and API to upload a time series, compare statistical and machine learning models and get forecasts.
FastAPI · Docker · ECS Fargate
Is your team facing a similar challenge?
Let's work togetherHow I can help
Data platforms
Pipelines and platforms that connect business data with its applications and analytics.
Data integration · Lakehouse · Pipelines
See related casePrediction and optimization
Forecasting and optimization engines in production, automated with MLOps.
Forecasting · Optimization · MLOps
See related caseData science and machine learning
Clustering, recommendation, classification and anomalies: models built for the problem.
Clustering · Recommendation · Anomalies
AI built into products and processes
Assistants and agents connected to your data, systems and processes.
Agents · RAG · MCP · APIs
See related caseArchitecture and production
Design, deployment and operation in the cloud, with CI/CD and infrastructure as code.
AWS · CI/CD · IaC · MLOps
See related case
Is your challenge outside these areas?Tell me about it
Solutions that run in the cloud
AWS is my primary platform. I also have experience with Azure Databricks, Microsoft Fabric and Azure Data Factory, and can evaluate other cloud options depending on the project.
- AWSPrimary platform
- Microsoft AzureEnterprise environments
- DatabricksLakehouse and ML
These names describe the platforms I work on; they imply no certification, partnership or endorsement by the providers.
From problem to operation
Understand
I clarify the need, the users, the constraints and how the result will be evaluated.
Design
I define the solution, the architecture and a scope that makes sense to build.
Build and deploy
I implement the components and integrate them with the environment where they will run.
Evaluate and improve
I review how the solution behaves and its results to guide the next iteration.
How I connect business, science and technology
I'm an industrial engineer with a specialization in IT project management and an international diploma in Business Administration. That background led me to work on operational and business problems, which I later connected with data science, forecasting and optimization, and with the engineering needed to turn models and data into products that run in the cloud.
I combine business sense, mathematical grounding, technology and creativity: a STEAM perspective I use to frame alternatives, design solutions and build them. I have worked with technical and business areas, from problem definition to production.
I have lived in three countries and worked with people from every continent. That multicultural experience helps me understand different contexts, communicate with diverse teams and shape each solution around the people who will use it.
I'm Diego Esteban Moreno Morales, known professionally as Diego Moreno.
Grounded decisions
Business and science: ask the right question, choose a relevant measure and check whether the model improves a decision.
Models that can run
Science and technology: turn models and methods into usable systems that can be integrated and evaluated in production.
Useful, viable products
Business and technology: design products that fit the processes, considering adoption, cost and viability.
At the center, the three dimensions together make it possible to build data and AI solutions for concrete problems and assess their contribution to the business.
Professional experience
Industries
- Retail and digital commerce
- Consumer goods and beverages
- Agribusiness and floriculture
- Financial services
- Steel industry
- Manufacturing
- Industry and logistics
Compass UOL
Data science and artificial intelligence
Data and AI solutions, architecture and development for enterprise projects across sectors.
PepsiCo, through Insight Global
Data science contractor
Working with the global team to improve forecasting in Latin America.
Florexpo
Forecasting
Forecasting for a floriculture company.
Frubana
Data science and demand planning
Forecasts and analytics to support demand planning.
Flower export group
Analytics and BI
Data and business intelligence to support operations.
Cencosud
Business analytics
Analysis and information tools for commercial and operational decisions.
Let's talk about your next challenge
If you need to build a data or AI solution, add experience to your team or discuss a professional opportunity, tell me what you have in mind.