Bonzanini Consulting

Senior ML and data science leadership, when you need it

I'm Marco Bonzanini, a London-based ML and data science consultant with 25+ years in software, data and machine learning.

I work with CTOs, heads of data, engineering leads and founders who need an experienced machine learning lead for a specific decision, a critical stretch of a project, or a team that is growing faster than its senior bench.

Get in touch See how I can help
Marco Bonzanini speaking at a conference
Marco Bonzanini, PhD

Services

Most engagements start with one of these queries. If yours doesn't fit neatly, get in touch anyway; plenty of good projects start somewhere in between.

  • ML Opportunity Assessment

    “Should we build this?”

    Before you commit a team and a budget, I look at the problem, the data you actually have and the constraints you're working under. You get a straight answer: build it, buy it, narrow the scope, or wait.

    • A written assessment with a clear recommendation
    • The main risks, and how to test them early and cheaply
    • A realistic plan and team shape if the answer is yes

  • Fractional ML Leadership

    “We need senior ML expertise.”

    A part-time head of ML or principal data scientist, for as long as you need one. I set technical direction, make the calls that need experience, and support the team day to day, so the work keeps moving while you hire or grow a permanent lead.

    • Technical direction, priorities and roadmap
    • Design, code and experiment reviews with the team
    • Hiring support: role definitions, interviews, onboarding

  • ML Technical Review and Advisory

    “Are we doing this right?”

    An independent look at a model, a pipeline or a whole ML platform. I review how the system is designed, evaluated and run in production, and tell you what is solid, what is fragile, and what to fix first.

    • Review of data, modelling, evaluation and deployment
    • Findings ranked by impact and effort
    • Follow-up advisory sessions if you want a sounding board

  • Training

    “We want to upskill our team.”

    Hands-on courses in Python, data analysis, software engineering and machine learning. I teach every course myself, and adapt each one to your team's level and the problems you work on.

    See the courses

Work

Selected engagements. Some clients prefer not to be named, so their projects are described by sector.

Online Car Rental

Recommendation Engine

The challenge
The client wanted to improve the relevance of offers presented to users searching for car rentals. The challenge was to determine how different signals (including price, ratings, popularity and user behaviour) should influence which offers appeared most prominently.
What I did
I developed and implemented a machine-learning ranking approach that combined behavioural and offer-level signals to improve the ordering of search results. This involved defining the ranking problem, identifying useful signals and designing an approach that could be integrated into the existing search platform.
Results
The new ranking approach produced more relevant search results and improved the quality of the offers presented to users, leading to revenue increases.

Recruitment

From research to production

The challenge
The client had strong research and engineering capabilities and was developing machine-learning solutions that showed promise in research and experimentation. They experienced frictions when turning those prototypes into systems that could be reliably deployed and operated in production.
What I did
I worked across the boundary between research and engineering, helping translate experimental approaches into production-ready systems. I advised on architecture, implementation and operational considerations, with a focus on making the resulting systems scalable, maintainable and practical to deploy.
Results
The client was able to take its machine-learning work from research prototypes into production systems and establish a more robust path from experimentation to deployment.

Banking and Finance

Training

The challenge
Multiple clients needed to train their team on machine learning and data science topics to improve their capabilities.
What I did
I designed and delivered custom training sessions tailored to their needs and technical level.
Results
The clients' teams gained valuable skills and knowledge, leading to improved performance and decision-making.

About

I'm an independent machine learning and data science consultant, and I work directly with the people who make technical decisions: CTOs, heads of data, engineering leads and founders.

My work sits between research, data science and software engineering. I'm most useful on projects where the difficult part isn't simply choosing a model, but working out how to formulate the problem, whether the available data can actually support it, and how to turn a promising approach into a production system.

My background is in search and ranking: I have a PhD in Information Retrieval, and I have since worked on data engineering and ML across sectors, including banking and finance, social media, flight safety, online retail, healthcare and recruitment.

I also teach and write about machine learning and data science: I'm the author of the book Mastering Social Media Mining with Python and several online courses published by Packt. I co-organised the PyData London meetup and conference series.

  • 25+ yearsExperience in software, data and machine learning
  • PhD in Information RetrievalSearch, ranking, decision models, evaluation
  • Author and Corporate TrainerPython, data analysis, machine learning
  • London and RemoteLondon-based; remote or on site
  • Co-organiser, PyData LondonMeetup and conference series

Training

Courses for teams working with Python and data. Durations are a starting point: I can shorten, combine or extend any of them.

All courses and details

Contact

Email is the quickest way to reach me. A few lines about your team, what you're working on and where you'd like help is plenty to start.

info@bonzaniniconsulting.com