O-Project by oaibuilds

Building modular intelligence — AI agents that don’t just support but operate, scale and create impact — Oriol Martinez

From Zero to Scalable Intelligence
— Starting at 30

🚀 Try O-Project Explore Projects Contact

A living lab of vertical AI agents built to read, decide, and operate.

Minimalism meets intelligence — building systems that think, adapt, and scale.

Python AI Agents Automation Backtesting Orchestration

About

Laying the groundwork for the AI frontier

I'm Oriol Martinez.

At 30, I made a pivotal career shift—starting from scratch to build a strong foundation in artificial intelligence and intelligent automation. Choosing depth over comfort, I’ve committed myself to developing the skills and mindset required for long-term impact.

I am now in my second year of an AI Engineering program at the Polytechnic University of Catalonia (UPC), where I focus on designing vertical AI agents and Python-based automation systems. My approach blends rigorous algorithmic thinking with hands-on execution, turning theory into functional, measurable solutions.

“Real learning begins when ideas come alive — Build → Test → Improve → Repeat.”

Each iteration brings me closer to scalable, intelligent systems that thrive in real-world conditions.

Beyond the technical journey, I’m also learning Mandarin to broaden my global perspective and build bridges across innovation ecosystems in Europe, Asia, and beyond.

再见，未来见 !

▸ Let’s connect or collaborate → oriol@oaibuilds.com

Oriol Martínez at Huazhong University of Science and Technology
Huazhong University of Science and Technology — Wuhan, China (2025)

    Projects

    The O-Project is an evolving ecosystem of vertical AI agents — currently at the prototype stage. Today, the tools focus on automation, structure, and orchestration without real AI reasoning, but the vision is to scale them into fully agentic systems that interoperate seamlessly across domains.

    O-Reader

    O-Reader is a document triage and routing tool. In its current version it uses light GPT-powered text analysis to classify and summarize PDFs, but the capabilities are intentionally simple. The roadmap is to evolve it into a richer semantic engine that can tag, enrich, and structure documents with contextual knowledge.

    View on GitHub →

    O-Trader

    O-Trader is a market automation and backtesting framework. At present it focuses on signal detection, risk control, and trade simulation via direct exchange APIs, without AI decision-making yet. The vision is to evolve it into an AI-assisted trading agent that adapts strategies and integrates with orchestration layers.

    View on GitHub →

    ⚠️ Current demo O-Reader runs with basic GPT integration but no autonomous AI decision-making yet. The vision is to evolve them into AI-driven agents that adapt dynamically, collaborate through orchestration layers, and scale into production-ready systems.

    Roadmap 2025–2027

    Now (2025)

    • O-Reader — PDF triage, routing & basic GPT summaries
    • O-Trader — market sandbox, signals & backtests (no autonomous AI yet)

    Next (2026)

    • O-Stream — content ingestion & publishing pipelines with prompt chains
    • O-Core — shared logic, orchestration & early multi-agent coordination

    Later (2027+)

    • O-Insight — dashboards for monitoring agent actions & fine-tuning in real time
    • O-Gateway — unified interface to connect external APIs & third-party services
    “We are on the verge of a transition — more important than the Industrial Revolution.”
    — Geoffrey Hinton     → Follow me on LinkedIn