Haydn Martin

Haydn Martin

Figuring out the best way to add value using data (then doing it).

Bio

Full-stack data scientist. Obsessed with how to leverage data to build cool and valuable things for people and organisations. Particularly interested in applications of theory, building things end-to-end, and doing stuff from scratch. Preferably projects that combine 2/3 of these things. Also financial data and FinTechs. I have 5+ years of experience working in finance at start-ups, scale-ups, and large institutions.

Timeline

2025 –

Senior Data Scientist at Elliptic

  • Data science and machine learning with blockchain data.

2023 – 2025

Senior Data Scientist at Tandem Bank

  • Created the Modelling & Prediction Team from scratch.
  • Responsible for team strategy, project selection, and ways of working.
  • Built three proprietary models with engineers and SMEs, now being productionised.
  • Using Python, SQL, TensorFlow, SageMaker, and more.

2021 – 2023

Data Insight Analyst at Tandem Bank

  • Led the bank's highest-priority loan data extraction and transformation project across 10+ data sources.
  • Pitched and scoped the bank's ML strategy; built the PySpark MLlib modelling framework from scratch.
  • Key role in post-acquisition Lakehouse integration.
  • Line-managed two data analysts.

2020

Risk Research Project at FNZ/Advicefront

  • Researched how financial advisers assess client risk tolerance and how to communicate investment risk to retail investors.
  • Designed a risk dashboard; used probability theory and Monte Carlo simulations in R.
  • Direct report to CEO.

2019

Growth Marketing Project at FNZ/Advicefront

  • Improved conversions from free-trial to paying customers using quantitative A/B testing and qualitative account-management techniques.

2019

Special Projects at Limitless

  • Jack-of-all-trades at a London FinTech start-up, reporting to the CIO.
  • Research, fundraising due diligence on 50+ investors, and competitor monitoring.

2017 – 2018

Collaboration Platform Manager at Principles for Responsible Investment

  • Managed PRI's Collaboration Platform — the world's leading proponent of responsible investing.
  • Maintained 100+ shareholder resolutions, distributed weekly client alerts, and produced analysis summaries.

2015 – 2016

Data Specialist at AXA Investment Managers

  • Part of the data team at Europe's largest real estate investment manager.
  • Organised financial and economic data for research analysts.
  • Built the Hotel Forecast database for improved forecasting.

Professional Reviews

His math and quantitative skills really helped us cut through the noise and focus on the main factors that drive investor behaviour and client-advisor communications. — Jose Supico, CEO, Founder
Haydn was a very effective colleague working with us on a range of diverse projects, often at very short notice, requiring his getting up to speed quickly both in terms of industry sub-vertical as well as in what we needed him to do within it. — Ka-ming Lim, Investor, CEO, Founder
Haydn is an exceptional manager with great leadership skills… has exceptional analytics and data science skills which he has utilised to great effect. — Christopher Scullion, Data Analyst
I strongly recommend him for analytical work in demanding environments. — Paolo Borella, Investor, VC

See more on LinkedIn

Projects

My public repositories focus on end-to-end machine learning. See GitHub for the full list. Examples:

bank-marketing-e2e-model — Creating a model from data in a UCI repo to web page UI to predict deposit-term subscribers from bank marketing efforts.

fraud-detection-advanced-algorithms — Comparing a neural network (in both TF and PyTorch) to XGBoost for predicting fraudulent credit card transactions.

crypto-virality-e2e-model — Extract features from crypto whitepapers and use that data to predict the market cap of the associated crypto product.

Writing

Professional

How Elliptic scales its intelligence without sacrificing its accuracy

Technical

Probably Not — Some random thoughts on data science things. Some examples:

Why ML in Finance Is Hard
Why and How I'm Organised
90% of Statistical Errors

Personal

Goodreads Book Reviews Narrative Non-fiction Personal Finance Paranoia

Academia

MSc Mathematical Finance, ISEG
Top 3 university in Portugal. Coursework in probability theory, financial instruments, stochastic calculus, and numerical methods in finance.

Master's dissertation: A Quantitative Investigation Into the Determinants of Risk Capacity

BSc Economics, University of Bath
Top 10 university in the UK. Coursework included statistics, data analysis, corporate finance, econometrics, investments, behavioural finance, etc.