Hi, I am Paul

Projects

Sudoku Solver
Sudoku Solver
Developer April 2020

Small serverless JS application that will solve all your sudokus by uploading a photo of your grid or manually enter each number.

Github
UN Migration Dashboard
UN Migration Dashboard
Developer March 2019

Shiny Dashboard displaying the poppulation migration flows as reported by the UN Department of Economic and Social Affairs.

Github

Recent Posts

Card image cap
Postgres Migrations

When aerich or alembic auto-generated migration files aren’t enough. Summary of the basic DDL/DML statements used to keep track of the Postrges DB migrations, followed by a couple of use cases. General DDL / DML CREATE / INSERT CREATE EXTENSION IF NOT EXISTS postgis; CREATE INDEX location_gidx ON product USING gist (location); CREATE UNIQUE INDEX product_name_supplier_uidx ON product (name, supplier); CREATE TYPE orderstatus AS ENUM('PROCESSED', 'SHIPPED', 'DELIVERED', 'CANCELLED'); -- upsert INSERT INTO product (name, supplier_id, location, status) VALUES ('NOTEBOOK-A6', 'Muji', ST_GeomFromText('POINT(12.

March 12, 2021 Read
Card image cap
Introduction to Colander

Despite being derived from Java Script, the JSON data format has quickly spread to other languages and python is no exception. Commonly referred to as dictionary, the simple key-value mapping offers a wide range of use cases: data storage, configuration and data transfer ! JSON serialization To make sure that the information contained in a dictionary corresponds exactly to the JSON open standard file format, we’ll need to serialise it.

August 12, 2020 Read
Card image cap
ISLR summary notes

This book describes all the basic ML models. Ideal to start a data science career when you haven’t got a degree in statistics or computer science. It is very comprehensible, has a lot of applications and exercises in R and is the companion book of the Stanford online course. Linear Regression Other good doc: sklearn StatsModels stats car (diagnostics) Classification Resampling Methods Linear Model Selection and Regulariation Beyond Linearity Ensemble Methods Support Vector Machines Unsupervised Learning

June 8, 2020 Read

Paul Martins

Data/ML Engineer at Deloitte, UK

I am a machine learning engineer at Deloitte where I help clients in their decision-making processes by building and delivering smart data products. My role mainly focuses on data transformation (data quality, ETL, performance optimization) and productionization of ML models.

Skills

Experiences

1
Machine Learning Engineer - Manager
Deloitte UK

May 2020 - Present, London, UK

The Risk Advisory team at Deloitte helps organisations manage their risk more effectively to unleash their full potential, creating and protecting value for all of their stakeholders.

Responsibilities:
  • Productionize the training and inference pipelines of a suite of XGBoost models for revenue and market penetration predictions
  • Built ETL jobs for near-real time data ingestion from several news providers
Data Science Consultant
QBiz Ltd.

May 2019 - May 2020, London, UK

QBiz is a London firm providing best-in-class consultants helping businesses address their analytical needs.

Responsibilities:
  • Help prototype and productionize a smart web-app for internet fibre network expansion
  • Participate in the development of a clustering module. Refactor code to decrease runtime
  • Code refactoring and features addition to a cost estimation module, decreasing runtime from hours to seconds
2
3
Data Scientist
Summit Media Ltd.

August 2017 - March 2019, London, UK

Summit is an award-winning digital marketing agency helping retailers transform their online business and grow their customer base.

Responsibilities:
  • Responsible of the R&D of Forecaster, a multi-channel marketing intelligence platform
  • Modularisation: refactor, package, test, document and dockerise legacy production code
  • Database migration testing
  • Development of a bid optimisation module to maximise online conversions (PPC and Shopping ads) for different audiences
  • Improve technical practices, continual learning and experimentation by organising / presenting at knowledge sharing sessions and supporting the team on coding best practices

Education

Queen Mary, University of London
2013-2017
Ph.D in Particle Physics
Activities
  • Analyse the data from the T2K experiment to extract relevant physics parameters
    • Events selection, Bayesian inference, statistical and systematic errors evaluation and propagation,
    • Quality control of the raw data recorded,
    • Theoretical model implementation in the Monte-Carlo events simulation,
    • Main analysis and data visualisation: C++, simulation in Fortran
  • Demonstrator and marker of courseworks for various undergraduate modules:
    • Mathematical Techniques, Nuclear Physics and Astrophysics, Elementary Particle Physics.
Paris-Saclay University
2011-2013
MSc in High-Energy Physics
Taken Courses
  • Quantum Field Theory
  • Particles and Symmetries
  • High-Energy Astrophysics
  • Strong and Weak Interactions
  • Detectors
  • General Relativity
  • Neutrino Physics
  • Quantum Mechanics
Extracurricular Activities
  • President of the Music Society