Skip to content

AI Consultant & Engineer

Agentic AI systems that make it to production

I help teams design, ship, and harden agentic workflows, hybrid RAG pipelines, and AI APIs that deliver measurable business results.

  • Deterministic orchestration for multi-step AI workflows and human-in-the-loop automation

  • Hybrid retrieval over OpenSearch, Qdrant, and PostgreSQL for higher-accuracy domain search

  • Production delivery with FastAPI, AWS, Azure, Docker, CI/CD, and clean architecture principles

  • Based in Spain and available for remote collaboration across Europe in English and Spanish

Book Free Intro Call Connect on LinkedIn

4+ years across AI engineering, data science, and data platforms. Azure and Databricks certified, with production systems serving daily workloads.

Portrait of Andrés Espinosa Sanfiel, AI Engineer

How I Help

I work best with teams that need more than a prototype. The goal is not just to demo AI, but to turn it into a maintainable system with clear business value and a realistic path to production.

  • Agentic AI Expert

    I design deterministic AI workflows with LangChain, PydanticAI, Agno, and the MCP protocol so multi-step systems stay reliable, testable, and useful in real operations.

  • RAG & Search Systems

    I build hybrid retrieval systems that combine vector and keyword search, improve domain accuracy, and give teams a cleaner path from document chaos to grounded answers.

  • Production AI Platforms

    I deliver AI backends that can actually support users: FastAPI services, streaming endpoints, cloud deployment, containerization, CI/CD, and architecture decisions that keep the stack evolvable.

Selected Outcomes

My focus is on systems that are useful, measurable, and maintainable. These are the kinds of outcomes I optimize for.

  • 70% Less Manual Processing

    Designed deterministic agentic workflows that reduced manual competitive analysis processing time by 70%.

  • 40% Better Retrieval Accuracy

    Built hybrid search pipelines over OpenSearch and Qdrant, improving domain-specific retrieval accuracy by 40%.

  • 500+ Daily API Requests

    Delivered production FastAPI services with real-time capabilities and daily usage in live environments.

  • Certified & Field-Tested

    Azure AI Engineer, Azure Data Scientist, Fabric Analytics, and Databricks certified, with hands-on delivery across AI engineering, data science, and data platforms.

Certifications & Recognition

I have restored the credential links so you can access the original certification pages directly from the site.

These case studies show how I approach automation, retrieval, and production delivery from business problem to architecture and measurable results.

  • Production AI API Platform

    Multi-step agentic workflows with deterministic orchestration, FastAPI, WebSockets, and AWS infrastructure. Result: 70% less manual processing and 500+ daily API requests served.

  • Hybrid Search RAG Pipeline

    A retrieval architecture combining OpenSearch and Qdrant to improve domain accuracy, support diverse query types, and keep LLM integration decoupled.

  • AI-Powered Email Automation System

    A production GenAI system that reduced daily email triage from 100+ items to 10-15 actionable items using Pydantic-AI, RAG, and FastAPI.

About Andrés

I bring a mix of economics, data engineering, data science, and production AI delivery. That combination helps me translate business requirements into architectures that are measurable, maintainable, and ready for real users.

Read About Me

Frequently Asked Questions

What technologies and frameworks do you specialize in?

My core stack includes Python, FastAPI, LangChain, PydanticAI, Agno, and the MCP protocol for agentic AI systems. For retrieval and data, I work with OpenSearch, Qdrant, PostgreSQL, PySpark, and SQL. I deploy across AWS and Azure with Docker, CI/CD, and MLOps practices when the use case needs them.

What kind of projects are you best suited for?

I am at my best when the challenge is moving an AI idea into a production-ready system. That includes agentic workflow automation, hybrid RAG and search, and AI APIs that need to serve real users with clear business outcomes.

How do you approach delivery?

I start from the business problem, define success metrics, and then design the system around reliability and maintainability. I favor deterministic workflows, clear interfaces, and clean architecture so the solution is easier to evaluate, test, and evolve over time.

Can you work with existing engineering teams?

Yes. I can work as an individual contributor, as a technical partner for an internal team, or as the person who turns an early proof of concept into a more robust architecture. I am comfortable collaborating with product, engineering, and data stakeholders.

Where are you available?

I'm based in Spain and open to remote work across Europe. I work in both English and Spanish and can support freelance, contract, and selected long-term engagements.

What sets you apart from other AI engineers?

My background in economics gives me a strong business perspective that complements the engineering work. I do not treat AI as a demo layer. I focus on systems that can be defended technically, measured operationally, and improved over time without rebuilding everything from scratch.

Start the Conversation

If you need help evaluating an AI opportunity, de-risking an architecture, or moving a prototype into production, I’m happy to discuss fit and next steps.

Book Free Intro Call LinkedIn