Gianmario Spacagna, Developer in Rome, Metropolitan City of Rome, Italy
Gianmario is available for hire
Hire Gianmario

Gianmario Spacagna

Software Developer

Rome, Metropolitan City of Rome, Italy

Toptal member since February 11, 2013

Bio

Gianmario is a hands-on AI leader with 13 years of professional experience and deep expertise in machine learning, data science, and AI product development. He has led teams and built scalable AI systems for startups and global companies alike. With a strong mix of technical depth and strategic vision, Gianmario helps organizations turn complex AI challenges into real, measurable business impact.

Portfolio

Moonsong Labs
Python, Large Language Models (LLMs), LangGraph, Google Cloud Platform (GCP)...
Brainly
Python, Deep Learning, Computer Vision, Natural Language Processing (NLP)...
Helixa
Python, Machine Learning, Deep Learning, Apache Spark...

Experience

  • Java - 4 years
  • Bash - 3 years
  • C - 2 years
  • C++ - 2 years
  • Scala - 2 years
  • PHP - 1 year
  • Erlang - 1 year
  • C# - 1 year

Preferred Environment

Git, Unix, Amazon Web Services (AWS), Google Cloud Platform (GCP), Python

The most amazing...

...product I've built is a shared memory infrastructure that enables coding agents to learn from each other, which led to spinning off The Memory Company.

Work Experience

Principal AI Engineer

2025 - 2025
Moonsong Labs
  • Led the research and development of Spark, an agentic shared memory system enabling coding agents to share experiential knowledge and foster active learning.
  • Architected Spark’s core components, spanning knowledge extraction and curation pipelines, agentic search (vector and graph-based), feedback loops, automated data seeding, and LLM-driven reasoning and coding capabilities.
  • Drove experiments, optimizations, and cloud deployment that brought Spark’s technical vision to life and led to the spin-out of MemCo.ai (The Memory Company).
  • Partnered with MemCo’s founding team (CEO and CTO) to define the product roadmap, technical infrastructure, and go-to-market and distribution strategy—supporting technical demos, fundraising, and partnerships.
Technologies: Python, Large Language Models (LLMs), LangGraph, Google Cloud Platform (GCP), MongoDB, OpenAI, OpenAI GPT-4 API, Claude API, Cursor AI, Tavily, Agentic AI

Director of Artificial Intelligence

2021 - 2025
Brainly
  • Partnered with the CTO to define and execute the company’s AI strategy, identifying technology opportunities that shaped product direction and business priorities.
  • Built and led the AI Services Department, scaling 40+ professionals across multiple teams, defining processes, culture, and career paths, and mentoring around 10 team leads to drive agility and high performance.
  • Delivered 25+ AI models to production, enhancing personalized learning experiences for over 350 million users worldwide.
  • Spearheaded Brainly’s MLOps platform and infrastructure, automating the machine learning lifecycle to improve efficiency, reliability, and scalability.
  • Unified research, engineering, and product efforts, providing technical leadership on POCs and production systems to ensure state-of-the-art AI integration across Brainly’s ecosystem.
Technologies: Python, Deep Learning, Computer Vision, Natural Language Processing (NLP), Strategy, Leadership, PyTorch

Chief Scientist, Head of AI

2018 - 2021
Helixa
  • Defined and executed the AI roadmap, aligned with company vision, customer needs, and financial goals, leading to a successful company acquisition.
  • Built and led a high-performing cross-functional AI team sourced from top 1% talents, coaching their career and mentoring them to excel in technical quality.
  • Developed proprietary AI algorithms that analyzed billions of social media interactions, uncovering statistically significant insights while preserving consumers privacy.
  • Reduced infrastructure costs by 90% by migrating Helixa’s legacy architecture to a serverless distributed system, delivering real-time insights in under 10 seconds and achieving previously unattainable performance levels.
  • Delivered compelling presentations on proprietary AI solutions, such as the Audience Projection algorithm, strengthening Helixa's reputation in the scientific and tech communities.
Technologies: Python, Machine Learning, Deep Learning, Apache Spark, Amazon Web Services (AWS), X (formerly Twitter), Serverless Architecture

Senior Data Scientist

2016 - 2018
Pirelli
  • Developed five ML models for connected vehicle products (Pirelli Connesso and Cyber™Fleet), enabling predictive insights into tyre wear and pressure, improving customer safety and efficiency.
  • Enhanced business decision-making by delivering granular monthly sales forecasts sliced to product segments and global markets.
  • Coordinated two innovative research projects with the Polytechnic of Milan, advancing AI-driven automotive solutions.
Technologies: Python, Scala, Scikit-learn, Forecasting, Regression, Internet of Things (IoT)

Data Scientist, AVP

2015 - 2016
Barclays
  • Designed predictive models for mortgage sales teams, boosting lead conversion rates through targeted recommendations.
  • Created algorithms for SmartBusiness Insights, transforming raw data into actionable dashboards for thousands of retail clients.
  • Developed a merchant entity resolution and record linkage system to normalize transactional data with local shops and branches.
Technologies: Scala, Apache Spark, Big Data, Advanced Analytics, Functional Programming

Data Scientist

2013 - 2014
Cisco
  • Pioneered data feed quality evaluations, ensuring the integrity and reliability of data used in analytics.
  • Developed advanced pattern recognition techniques for identifying and blocking malicious web domains.
  • Enhanced cybersecurity measures by integrating data-driven insights to proactively detect and prevent harmful web threats.
Technologies: Apache Pig, Scalding, Apache Hive, HBase, MapR, Hadoop

Big Data Software Engineer Intern

2012 - 2013
AgilOne
  • Developed an R&D project relating to parallel genetic auto-tuning systems for big data machine learning algorithms ("TunUp").
  • Studied state-of-the-art techniques for model fitting and machine learning.
  • Tested and evaluated performance of algorithms over many datasets and compared said performance with other non-heuristic techniques.
  • Designed all solutions to exploit parallelism using the Amazon cloud infrastructure.
Technologies: Machine Learning, Java, Big Data

gTech Engineer Intern

2012 - 2012
Google
  • Implemented end-to-end features aimed at improving internal reporting and data analysis platform.
  • Translated customer requirements into technical specifications.
  • Integrated code and re-factored architecture of existing complex systems.
  • Submitted 14,000 lines of reviewed Java code into production, including 35% back-end services and 65% UI based.
  • Designed and developed a web user interface for the aforementioned features.
  • Designed data model and implemented physical representation using BigTable.
Technologies: EasyMock, Guice, BigTable, CSS, Protocol Buffers, Git, GWT, Java

Experience

Spark

https://www.memco.ai/
A plug-in memory layer for AI IDEs like Cursor. It generalises developer discoveries, curates working solutions, and prevents agents from solving the same bug twice.
Works with all major IDEs as an MCP
Seamlessly integrate with your favorite development tools.

Education

2019 - 2020

Master of Business Administration - MBA in Business Administration and Management, General

Quantic School of Business and Technology - Washington DC

2011 - 2013

Master's Degree in Software Engineering of Distributed Systems

KTH Royal Institute of Technology - Stockholm, Sweden

2010 - 2013

Master's Degree in Computer and Communication Networks Engineering

Polytechnic of Turin - Turin, Italy

2007 - 2010

Bachelor's Degree in Telecommunications Engineering

University of Cassino - Cassino, Italy

Skills

Libraries/APIs

Scalding, EasyMock, X (formerly Twitter) API, Facebook API, jQuery, Claude API, PyTorch, Scikit-learn

Tools

MapR Tables, Tableau, Adobe Fireworks, Git, Adobe Photoshop, MATLAB

Languages

Scala, Java, C, C++, CSS, PHP, Python, Erlang, Bash, C#, Ruby

Frameworks

GWT, Swing, Guice, Ruby on Rails (RoR), JUnit, Hadoop, Google Protocol Buffer, JavaServer Faces, JPA, LangGraph, Apache Spark

Storage

Apache Hive, HBase, BigTable, NoSQL, MySQL, MongoDB

Paradigms

Functional Programming, Object-oriented Programming (OOP), Agile Software Development, Constraint Programming, Imperative Programming, Serverless Architecture

Platforms

Linux, MapR, Apache Pig, Unix, Eclipse, Drupal, Google Cloud Platform (GCP), Amazon Web Services (AWS), X (formerly Twitter)

Industry Expertise

Telecommunications

Other

Protocol Buffers, Machine Learning, Watchmaking Industry, Ajax, Strategic Planning, Data-driven decisions, Leadership, Software Engineering, Distributed Systems, Computer Networking, Large Language Models (LLMs), OpenAI, OpenAI GPT-4 API, Cursor AI, Tavily, Agentic AI, Deep Learning, Computer Vision, Natural Language Processing (NLP), Strategy, Big Data, Information Retrieval, Forecasting, Regression, Internet of Things (IoT), Advanced Analytics

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring