Presentation - Albert Masoliver's learning site

# Albert Masoliver **Position:** **CTO** at [GLAM SOFTWARE](https://glamsw.com/en) · **AI Systems & Distributed Architecture** **Location:** Girona, Spain ## Introduction I'm Albert Masoliver — a physicist by training who works at the intersection of **artificial intelligence** and **distributed-systems architecture**. My focus is the orchestration layer: designing the agentic platforms, verification pipelines, and memory architectures that make AI systems trustworthy enough to ship from, and the distributed-systems backbone that lets them operate at scale. My professional practice and study span four reinforcing tracks: ### 1. Artificial Intelligence — the full breadth I work across the field, from the symbolic foundations to modern generative systems: - **Classical AI** — search (A*, IDA*, MCTS), logic and knowledge representation, constraint satisfaction, planning (STRIPS, PDDL, HTN), probabilistic reasoning (Bayesian networks, HMMs). Index: [[7 - Classical AI Notes Hub]]. - **Machine Learning** — supervised, unsupervised, reinforcement learning; the full evaluation toolkit; feature engineering; ensemble methods (Random Forest, XGBoost) that still dominate tabular problems. Index: [[8 - Machine Learning Notes Hub]]. - **Deep Learning** — from the perceptron foundations up to convolutional, recurrent, and generative architectures (autoencoders, VAEs, GANs). Index: [[9 - Deep Learning Notes Hub]]. - **Optimization** — continuous, linear, convex, combinatorial, and metaheuristic methods. The substrate underneath all of ML and most of OR. Index: [[10 - Optimization Notes Hub]]. ### 2. Modern AI Software Engineering Beyond the foundations, my current operational focus is **agentic AI software engineering** — taking frontier LLMs from "interesting demo" to "load-bearing component of a production system": - LLM internals — tokens, context, reasoning budgets, the cost-quality fleet. - Agent patterns — tool use, ReAct, multi-agent pipelines, retrieval-augmented generation, prompt-injection threat models. - Orchestration practice — Spec-Driven Development, the Claude Code / MCP / hooks stack, memory architectures that survive sessions and teammates, DORA-aligned metrics that prove it's working. Distilled into a working course / playbook: [[0 - Modern AI Software Engineering Hub]]. ### 3. Distributed Systems & Domain Design My foundation as a software architect. I build production systems with **Domain-Driven Design** and **EventStorming**, applying tactical patterns (aggregates, value objects, domain events) within strategic context maps. Index: [[11 - DDD Notes Hub]]. ### 4. Databases I still take databases seriously — query optimisation and indexing are where many distributed systems live or die. Index: [[12 - Query Optimization and Indexing Notes Hub]]. --- My through-line: I'm convinced the next decade of software engineering is shaped by people who treat LLMs as **components** — priced, fallible, but composable — embedded inside soundly-architected systems. That belief shapes how I architect, build, study, and teach. Outside of work: **astronomy and astrophotography**. Exploring the universe both scientifically and artistically remains the thing I do for no professional reason at all 🚀🔭 ## Professional Experience ### GLAM SOFTWARE **CTO** *March 2016 - Present (10+ years)* I lead the technical strategy at GLAM — currently focused on integrating agentic AI capabilities into our products, while keeping the distributed-systems backbone (DDD-aligned, event-driven) sound enough to support that integration. Architecture decisions, agentic workflow design, code review, hiring, and the harder conversations with stakeholders about what AI actually can and can't do. ### BIDATA SOLUTIONS & CONSULTING, S.L. **CTO** *September 2016 - May 2018 (1 year 9 months)* Directed the technological operations in Girona, ensuring successful implementation of innovative data and analytics solutions. ### Just Enginyeria de la Informació **Partner** *March 2001 - February 2016 (15 years)* Co-founded the company and specialised in information engineering solutions in Banyoles, Spain. ## Current Work - Authoring **Modern AI Software Engineering: The Orchestrator's Playbook** — a seven-module practitioner's course on building, governing, and verifying agentic AI workflows in production codebases. Covers the thinking economy (tokens, reasoning budgets, model selection), Spec-Driven Development, agentic CLIs, the extension trifecta (skills / MCP / hooks), memory orchestration, the meta-agent factory & verification frontier, and the human-in-the-loop ROI lens. - Maintaining an open, interlinked **personal knowledge base** of 250+ atomic notes spanning Classical AI, Machine Learning, Deep Learning, Optimization, LLM internals, agentic patterns, RAG, orchestration practice, Domain-Driven Design, and database internals. The substrate for both the course and the architectural work at GLAM. ## Education - **Universitat de Barcelona** Bachelor's Degree in Physics, Theoretical Physics *1993 - 1997* - **Universitat Oberta de Catalunya** Applied Data Science Degree *2020 - 2024* - **Universitat Oberta de Catalunya** Postgraduate Diploma in Business Intelligence and Data Analysis *2016 - 2017* - **Universitat de Barcelona** Engineer's Degree in Electronic Engineering *1996 - 1999* ## Skills - **Languages:** Catalan (Native), Spanish (Native), English (Professional), French (Professional) - **Artificial Intelligence:** Classical AI (search, logic, planning, probabilistic reasoning), Machine Learning (supervised / unsupervised / reinforcement), Deep Learning (CNNs, RNNs, transformers, autoencoders, GANs), Optimization (continuous, linear, convex, combinatorial, metaheuristic). - **Modern AI Engineering:** Agentic CLI tooling (Claude Code, OpenCode), Model Context Protocol (MCP), Spec-Driven Development (OpenSpec, BMAD), prompt caching, reasoning-budget tuning, retrieval-augmented generation, multi-agent verification pipelines, memory orchestration (ADRs, layered memory, auto-memory). - **Architecture & Systems:** Domain-Driven Design, EventStorming, Event Sourcing & CQRS, distributed-systems patterns, query optimisation and indexing. - **Languages & Stacks:** C#, Python, JavaScript / TypeScript, Microsoft SQL Server. - **Foundations:** Physics, applied mathematics, statistics — bridged forward into the AI / ML stack. ## Certifications - Introduction to AWS - Discrete Optimization ## Contact - **Email:** [[email protected]](mailto:[email protected]) - **LinkedIn:** [linkedin.com/in/albert-masoliver-56882b12](https://www.linkedin.com/in/albert-masoliver-56882b12) - **Website:** [glamsw.com](http://www.glamsw.com/)