Personal Data

Work Experience

2021-20XX

Data Engineer: Competition and Markets Authority

I moved to the CMA in Spring 2021 for a IC (Individual Contributor) role, concentrating on productionising large scale machine learning models and building data pipelines to support the department’s digital forensics and market investigation activities. The role blends DevOps, Data Engineering, and Data Science skills, and particularly focuses on how to apply these skills to solve the department’s business problems in multidisciplinary teams of SMEs, technical leads and product managers. We work within the AWS ecosystem.

2020-2021

Head of Data Science Unit: HM Treasury

In February 2020, I was transferred to the UK’s Finance and Economics Ministry. I built a team around me to introduce Data Science methods in its policymaking process for the first time. Utilising skills in Data Engineering, Data Science and Leadership, I have quickly ended up key to the department’s response to the COVID-19 pandemic; setting the department’s strategy for the adoption of Data Science as an analytical discipline, leading projects designed to bolster the evidence base for government support schemes, and coordinating colleagues from other government departments to ensure efficient project delivery under extreme time constraints.

2019-2020

Data Science Team Leader: Department for Business, Energy and Industrial Strategy

I led a team of Data Scientists working in Rapid Response and advanced techniques at the department. I was the senior official in charge of high-profile projects involving up to 6 other data scientists working concurrently. I also used my software engineering skills to modernise the team’s Data Science process.

2018-2019

Data Scientist: Department for Business, Energy and Industrial Strategy

I moved to BEIS to work in their Data Science team, particularly on EU Exit Analysis and Rapid Response projects for ministers and the Permanent Secretary. My projects were heavily focused on NLP, Dataset Creation & Webscraping, Interactive Reporting and Classification models.

2017-2018

Data Scientist: Food Standards Agency

I joined the Civil Service Fast Stream and was placed at the Food Standards Agency. I developed the Data Science capability at the department.

My projects included world trade visualisation, and predicting Norovirus frequencies as a function of Tweets with >80% accuracy.

2015-2017

Consultant: InfoCat Ltd.

My role was to design, build and integrate financial systems. I worked in a wide variety of sectors, including Investment Banking, FMCG, Charity, Travel and Aerospace firms.

I learned a wide range of Data Engineering skills, as the most challenging parts of the role were gluing different systems together in an efficient way. Unlike most consultancies, we had a reputation for our focus on resilience - I learned how to write fantastic documentation and robust software here.

Education

2011-2015:

MSci Mathematics (Geophysical and Astrophysical Fluid Dynamics): First Class

Research-focused Masters intended to prepare students for PhD study. My research was in Geophysical Fluid Dynamics (cyclone modelling).

My taught courses were a fairly even mix of Calculus, Dynamical Systems, Climate Science and Linear Algebra. Details on these courses, and my marks on them, are available in my degree transcript on request.

Skills

Technical

Python

I am an expert Python user for all data tasks. I have used it for ETL/Data Pipelines, Deep Learning, Web Applications, NLP and Package Development.

R

I am an expert R user. I have used it mostly for visualisations, ETL, statistical modelling, forecasting and interactive reports. I have no preference for either Python nor R generally speaking; I am language agnostic and use the best tool for the job given the wider circumstances of the project.

SQL

I use SQL to write performant ETL production code. I have used both traditional row-indexed DBs like SQLite, Postgres & SQL Server, and columnar DBs like DuckDB and BigQuery for analytics workloads, both in production.

DevOps

I manage cloud infrastructure in the AWS Ecosystem primarily, but I have experience with Azure. I manage the infrastructure using the AWS CDK & CloudFormation. My projects use CI/CD, many employ containers (particularly those with complex system-level dependencies or serverless). I’ve deployed projects that employ monolithic and serverless architectures, and use whichever makes the most sense for the project (cost, BAU plan, project lifetime, dev team structure).

Git

I am an advanced Git user. I use Magit to manage my repositories. I also implement QA processes in my teams in part with a branch + PR workflow.

Spark

I have used Spark (via the PySpark API) to filter and aggregate large datasets held within a Data Lake on our production cloud infrastructure.

Linux

I am comfortable using UNIX-based OSes both via SSH and with a Desktop Environment. My favoured development environment is Ubuntu with Emacs, which I make work either with native linux or WSL2.

LaTeX

I use LaTeX for document authoring, including this CV.

Soft Skills

• Team Leadership & ability to set analytical strategy.
• Rigorous Learning & Development focus for myself and my team.
• Excellent communicative skills with stakeholders.
• Excellent track record of translating business needs into Data Science products.
• Product Ownership, Planning & Delivery to production.

Interests

Sports

Formula 1 & Football.

Cooking

I primarily cook Italian, Greek, Chinese and Japanese dishes. I am also a keen baker.

Books

I read non-fiction primarily. Fiction books are generally Sci-Fi. My favourite book is Musashi, by Eiji Yoshikawa, a fictionalised depiction of Miyamoto Musashi’s life.

Projects

Here are two toy projects I have worked on in my spare time. They are not designed to be a showcase of my technical expertise, which has mostly been built through delivery of closed-source projects in government.

TF-IDF String Matcher

https://github.com/LouisTsiattalou/tfidf_matcher

Python package providing lightning fast string matching using TF-IDF & K-NN clustering on ngrams (default 3-grams) of strings. \(\mathcal{O}(n\log{}n)\) complexity.

Trade Data API

https://github.com/LouisTsiattalou/TradeDataAPI

HMRC Trade Data is extremely inaccessible in format. This project aims to download and ingest the dirty data into a consistent and organised database, with an API layer on top.

Current Status: ETL processes complete. Project on hold as all my time is currently spent working on the COVID Response.