ARC

Andrés Ruiz Calvo

Madrid, SpainEmailLinkedInGitHubPersonal WebsiteAd Astra AI BlogYouTube

AI Engineer shipping production GenAI for enterprise customers. Customer-facing technical advisor—trained 100+ engineers on GenAI and RAG. MSc in AI/ML; University ML Professor at UAX. Founder of Ad Astra AI — a technical hub for creating open-source production-ready AI projects to technical and non-technical audiences (blog, YouTube, open-source GenAI). Databricks (Data Engineer & GenAI Engineer), 3x AWS, and 3x Anthropic certified.

Work Experience

IBM

Madrid, Spain

AI Engineer

Nov 2024 - Present
  • Shipped production RAG to 100+ enterprise users: Elasticsearch multi-stage retrieval, cross-encoder re-ranking, and offline evaluation of chunking/indexing before go-live; event-driven ingestion and containerised FastAPI services kept knowledge bases current and reliable in production.
  • Gave client teams day-to-day access to proprietary documentation via AI assistants and IDE integrations (MCP servers), reducing friction for engineers querying complex content.
  • Cut developer time by 30% with multi-agent GenAI workflows for code generation, self-correction, and unit testing—shipped from prototype to production on cloud infrastructure.
  • Trained 100+ customer engineers as technical advisor, turning GenAI designs into patterns teams could adopt in production.
  • Led autonomous-agent POC for senior leadership and customer stakeholders, securing buy-in for next-step adoption.

UAX University

Madrid, Spain

University Professor in Machine Learning

Oct 2025 - Present
  • Associate Professor of Machine Learning I & II in the AI and Computation degree (20+ students): own and deliver both courses end-to-end—syllabus, lectures, labs, and assessment—covering supervised and unsupervised learning, model evaluation, and applied ML workflows.

Huawei

Madrid, Spain

Junior Solutions Architect

Nov 2023 - Nov 2024
  • Closed hybrid-cloud engagements by translating stakeholder requirements into approved architectures with clear cost and operational trade-offs.
  • Migrated startup AI workloads to public cloud, clearing infrastructure bottlenecks so the team could run inference and serve assets reliably from a single footprint.
  • Productionised ML inference for a software company under bursty load—autoscaling and queueing so jobs completed reliably during traffic spikes.
  • Delivered backup solution for a telecommunications consulting firm: defined recovery objectives and paired technical design with economic proposals for stakeholder approval.

Education

Master's Degree in Machine Learning and Healthcare

2022 - 2023 | Madrid, Spain

Universidad Carlos III de Madrid

Thesis: Applied supervised machine learning to estimate the best intraocular lens (IOL) for implantation after cataract surgery from pre-operative images; up to 11% improvement over prior work, in collaboration with CSIC.

Bachelor's Degree in Data Science and Engineering

2018 - 2022 | Madrid, Spain

Universidad Carlos III de Madrid

Thesis: Built a deep learning semantic segmentation model in PyTorch on real-world post-cataract patient eye images; 7% improvement over state of the art for cataract and presbyopia assessment, in collaboration with CSIC.

Skills

Data & ML platforms: Databricks, PySpark, Delta Lake, Spark Structured Streaming, Medallion architecture (Bronze/Silver/Gold), Auto Loader, Spark MLlib

AI / ML: LLMs, RAG, multi-agent workflows, LLM fine-tuning (LoRA, SFT), vector search, embeddings, reranking, GenAI evaluation, prompt engineering, deep learning

Frameworks & Libraries: LangGraph, LangChain, HuggingFace, FastAPI, PyTorch, pandas, scikit-learn, Unsloth, LiteLLM, Crawl4AI, Docling

Languages: Python, TypeScript, SQL

Cloud & DevOps: AWS (ECS Fargate), Docker, Jenkins CI/CD, Red Hat OpenShift, GitHub Actions, Prometheus, Grafana

Projects

Satellite Constellation Predictive ML Maintenance System (Databricks)

Built an end-to-end ML pipeline on Databricks: PySpark and Delta Lake with Spark Structured Streaming for real-time satellite telemetry; Medallion architecture (Bronze/Silver/Gold), Auto Loader ingestion, schema evolution, and Spark MLlib models for predictive maintenance.

End-to-end Agentic RAG deployed on AWS (ESA Sentinel documentation)

Developed a production-ready AI-powered RAG chatbot deployed on AWS ECS Fargate for ESA Sentinel mission documentation using a Python/FastAPI backend with LangGraph agentic workflows and Qdrant vector database. Built a Crawl4AI-based web crawler for automated ingestion from SentiWiki (200+ documents) with real-time streaming, query decomposition, and hybrid search with reranking.

GitHub: sentiwiki-ai

LLM fine-tuning on space exploration standards (ECSS)

Built an end-to-end pipeline from licensed ECSS standards PDFs to instruction tuning: Docling-based extraction to Markdown, deterministic SFT construction, and optional higher-quality supervision via an external LLM (Anthropic). Fine-tunes a small language model with Unsloth (LoRA / 4-bit) for deployment on GPU cloud (RunPod), with containerised training via Docker.

F1 AI Race Engineer — Generative AI on real Formula 1 data

Built a GenAI product for F1 race analysis: users load a completed Grand Prix from FastF1 timing/telemetry exports, then run LLM-generated race narratives (summary, lap-by-lap story beats, chapters), interactive Q&A on drivers and strategy, and replay-aware lap context.

GitHub: f1-ai-race-engineer

Certifications

Claude Certified Architect — Professional — Advanced architecture for production Claude systems — agent design, tool orchestration, evaluation, and enterprise deployment patterns.

Claude Certified Developer — Foundations — Foundational skills for building applications with Claude — API integration, prompting, and developer workflows.

Claude Certified Architect — Foundations — Architecture patterns, tool use, and production-minded agent design with Claude.

Databricks Certified Generative AI Engineer — Associate — LLM-enabled solutions on Databricks: RAG, Vector Search, Model Serving, MLflow, Unity Catalog.

Databricks Data Engineer — Associate — Batch/streaming pipelines, Delta Lake, warehouses, lakes, ETL, PySpark (Databricks Intelligence Platform).

AWS Certified Machine Learning — Specialty — Building, training, tuning, and deploying ML on AWS (two years' experience validated).

AWS Solutions Architect — Associate — Secure, scalable, cost-optimised designs across AWS services.

AWS Certified AI Practitioner — Foundational AI/ML and AWS implementation practices.

Languages

Spanish — NativeEnglish — C1French — B2

Workshops

PyConES — Multimodal RAG (50+ participants)

Hands-on workshop at PyConES (Spain's largest Python conference): instructed 50+ participants on building multimodal RAG from scratch with Python and IBM Docling; presented technical concepts to mixed technical audiences.

Workshop link

Ad Astra AI & Open Source

Ad Astra AI — blog, YouTube, and open-source GenAI

Created Ad Astra AI, a technical hub where I build open-source production-ready AI engineering projects: multi-part technical series (agentic RAG on AWS, evaluation, deployment) written for engineers who build AI projects and curious non-technical readers alike. Publishes open-source reference implementations on GitHub (SentiWiki-AI, F1 AI Race Engineer).

Ad Astra AIGitHubYouTubeSubstack

Research Publications

Automatic segmentation and quantification of OCT images before and after cataract surgery using deep learning

First author, Applied Optics: developed in Python a deep learning segmentation algorithm (Feature Pyramid Network with pre-trained encoder) for anterior-segment OCT images, achieving 93.2% accuracy and 0.34 s/image processing.

Publication link

Use the download button or press Cmd/Ctrl + P to save or print this CV