Summary

I specialize in building and managing algorithms to optimise Pay-Per-Click (PPC) spend, influencing a significant yearly budget of $1 billion. My role encompasses a wide range of responsibilities, from model building, conducting and analyzing A/B tests to deep data mining and visualization.

I work closely with a diverse team, using my technical skills to drive projects, guide junior members, and collaborate with project management to align our efforts with the business's priorities.

My role is integral in maintaining the balance between stakeholder needs and technical execution, owning end-to-end processes, and innovating through exploratory data analysis.

In this role, I used in Python, SQL, A/B testing, statistics, non-linear optimisation, machine learning, software engineering, CI/CD, OOP, data analysis, and data engineering.

Key Achievements

• Non-linear Optimisation - Built and managed a non-linear optimisation algorithm to estimate elasticity, and maximise GP for a given level of spend and constraints, using optimisation package CasADI, managing $1bn in marketing spend
• Reinforcement Learning - Deployed a bandit model to optimise campaign parameters, resulting in additional £10m in revenue
• Spend Prediction Model - Built a bid strategy spend prediction model that resulted in bidding frequency increase of 40% - from 4.3 times a month (once a week) to 6 (once every 5 days)
• Software Engineering Standards - Implemented software engineering best practices in our team, including containerisation, CI/CD, linting, unit testing
• YahooJapan - Built a CPC algorithm that managed all of YahooJapan spend

Key Responsibilities:

• Algorithm Development
• A/B Testing
• Productionising Algorithms and Systems
• Data Pipelines
• Internal Tooling and Infrastructure

Summary

I have been researching and building a model capable of generating flow rate graphs from the sound of male urine voiding into a toilet.

These flow graphs are intended to be used, and interpretted, by medical professionals to assist in the diagnosis of urinary and prostate problems

This involves extensive research, algorithm development, and the process of bringing the model to a production-ready state

Using audio analysis, audio feature extraction and Causal CNNs, the project aims to map the acoustics of fluid dynamics into accurate flow graphs. This innovation has potential applications in healthcare for monitoring urinary health, offering a non-invasive and privacy-respecting method for collecting flow data.

My role in this project encompasses the entire lifecycle of the model - from initial concept and research to algorithm design and final productionisation.

I also built the website (SaaS) using React and Django, as well as the App using Flutter

Key Achievements

• Inference model - Engineered a novel Convolutional Neural Network architecture for the core inference model that outperforms LSTMs at the task, achieving a RMSE of 3ml/s accuracy
• Dereverberation - A model to remove early and late reflections from audio, a significant problem when recording in a confined spaces of differing geometries and building materials to process the audio to create room invariant recordings. Built an informed dereverberation model using U-Net architecture and a novel cross-attention mechanism which reduced reverb artifacts by 80%
• Detection models - Created flow detection and anomaly detection algorithms, enhancing the accuracy and reliability of the inference process.
• Data Alignment - Developed novel algorithms for parsing, cleaning, and aligning input and ground truth data, ensuring high-quality data processing and accuracy for model training
• Full Stack - Built the frontend (React), backend (Django), and App (Flutter) infrastructure.

Key Responsibilities:

• Model Research and Development
• Data Parsing and Cleaning
• CI/CD development
• Platform development

Summary

I led the Data Innovation team of Data Engineers and Scientists at Samsung Ads for all of their European operations. We were focused on developing new data products and tools to understand our users, increasing data availability at scale amongst the Analytics team, and automation.

Key Achievements

• Powerpoint Automation - Designed and developed an internal Python tool, Nova, which could automate Powerpoint reports with up-to-date data, and be able to be used by non-technical teams. Adopted by the entire Analytics team for generating all monthly and quarterly reports
• OTTy Awards - Runner up for both Innovator and Problem Solver awards across all 3 business segments of Samsung Ads EU

Key Responsibilities:

• Productionising Models
• Managing Cloud Environments
• Internal Tool Development
• Platform Development

Summary

I joined in the early days of the analytics team where I built tools to automate and scale key deliverables for analytics, built and optimised queries, developed cloud infrastructure and data pipelines. I continued this role throughout studying my MSc.

Key Achievements

• Demographic Prediction Model - Productionised a model used to predict demographics based on linear, streaming and gaming viewing behaviour
• Self-service reporting - Built front and back-end of an internal self-service reporting platform Express Analytics, using Django. This ended up saving the analytics department 4 full-time analysts worth of time and was duplicated by the US Analytics team
• Data management tool - Built internal python package, Eclipse, used to simplify and manage all database, S3, query manipulation, GitHub operations. Now standard across all python projects across EU Analytics. This has also been adopted by the LATAM and APAC Analytics teams.
• Manage Cloud Environments - Manage AWS environment and assets for all of EU.

Key Responsibilities:

• Productionising Models
• Managing Cloud Environments
• Internal Tool Development
• Platform Development

Summary

As a result of the automation and innovation I’d brought to the Insight department I was promoted to Data Scientist after 7 months. Here I focused on leveraging Dixon’s data and cloud environments to predict and respond to customer behaviour.

Key Achievements

• Propensity Modelling - Built and productionised a machine learning model to predict a customer’s likelihood to purchase. Identified a segment who are 9 times more likely to convert
• Governance - – Created and implemented policies on the production and maintenance of code in projects, as well as governance of GCP projects
• Database Migration - Migrated reporting tables and processes from on-premise databases to GCP as part of wider systems automation

Key Responsibilities:

• Productionising Models
• Managing Cloud Environments
• Internal Tool Development
• Platform Development

Summary

I worked with the online marketing channels and digital campaign teams to analyse effectiveness of marketing campaigns and website personalisation projects

Key Achievements

• Auditing - Saved £39k/month after investigating discrepancy between 2 revenue reporting systems. I built a Python script that verified transactions by combining cloud and on-premise data
• A/B testing - Created a series of SQL queries that determined statistical significance of A/B test metrics that wasn’t possible with existing tools
• Automation - Automated reporting processes using Python and SQL, saving the Insight team 8 hours per week through efficiency and reduction in errors
• Consistent high achiever - A consistent track record of going the extra mile, including volunteering for database migrations, presenting to students at Edinburgh University on behalf of Dixons Carphone, and creating/running training sessions for other internal departments

Key Responsibilities:

• Productionising Models
• Managing Cloud Environments
• Internal Tool Development
• Platform Development

Summary

I studied my MSc Artificial Intelligence full-time during COVID whilst working for Samsung. I gained a distinction with an average mark of 79%.

Topics

• Machine Learning - Clustering, Decision Trees, Regression, SVM, KNN
• Computer Vision - Image Processing, Feature Detection, Segmentation, Motion, Recognition
• Biologically Inspired Methods - Binary and Continuous Genetic Algorithms, Ant Colony Optimisation, Particle Swarm Optimisation
• Reinforcement Learning - MDPs, Thompson Sampling, Q Learning, Temporal Difference Learning
• AI Planning - Heuristic Planning, Partial Order Planning, Temporal and Resource-based Planning, MDPs and Partialy Observable MDPs, Hierarchical Task Network

I studied the Chartered Financial Analyst (CFA) exams while actively engaged in my professional responsibilities. Below are the topics covered in the Exam:

• Ethical and Professional Standards: Studied principles and guidelines governing ethical behavior in the investment profession.
• Quantitative Methods: Explored statistical techniques and methods for analyzing financial data.
• Economics: Examined macroeconomic and microeconomic concepts relevant to investment decision-making.
• Financial Reporting and Analysis: Analyzed financial statements to assess the financial health and performance of companies.
• Corporate Finance: Learned about capital budgeting, cost of capital, and corporate governance.
• Equity Investments: Explored valuation methods and strategies for investing in stocks.
• Fixed Income: Studied bond markets, interest rate risk, and fixed income securities.
• Derivatives: Explored options, futures, and other derivative instruments.
• Alternative Investments: Learned about hedge funds, private equity, and other alternative investment vehicles.
• Portfolio Management: Studied portfolio theory, asset allocation, and portfolio construction techniques.

I studied the Chartered Financial Analyst (CFA) exams while actively engaged in my professional responsibilities. Below are the topics covered in the Exam:

• Ethical and Professional Standards: Studied principles and guidelines governing ethical behavior in the investment profession.
• Quantitative Methods: Explored statistical techniques and methods for analyzing financial data.
• Economics: Examined macroeconomic and microeconomic concepts relevant to investment decision-making.
• Financial Reporting and Analysis: Analyzed financial statements to assess the financial health and performance of companies.
• Corporate Finance: Learned about capital budgeting, cost of capital, and corporate governance.
• Equity Investments: Explored valuation methods and strategies for investing in stocks.
• Fixed Income: Studied bond markets, interest rate risk, and fixed income securities.
• Derivatives: Explored options, futures, and other derivative instruments.
• Alternative Investments: Learned about hedge funds, private equity, and other alternative investment vehicles.
• Portfolio Management: Studied portfolio theory, asset allocation, and portfolio construction techniques.

During my civil engineering degree, I covered the following topics:

• Mathematics: Studied calculus, differential equations, and linear algebra applied to engineering problems.
• Physics: Explored principles of mechanics, thermodynamics, and electromagnetism relevant to engineering applications.
• Statics and Dynamics: Analyzed the behavior of structures and systems under static and dynamic loads.
• Structural Analysis: Learned methods for analyzing the internal forces and stresses in structures.
• Geotechnical Engineering: Studied soil mechanics, foundation design, and earthwork construction.
• Hydraulics and Hydrology: Explored fluid mechanics, open channel flow, and hydrological processes.
• Transportation Engineering: Learned about transportation systems, traffic flow, and pavement design.
• Environmental Engineering: Examined water and wastewater treatment, air pollution control, and solid waste management.
• Construction Management: Studied project planning, scheduling, cost estimation, and construction methods.
• Engineering Ethics and Professionalism: Explored ethical considerations and professional responsibilities in civil engineering practice.