Neil Mcblane

Architecting, developing, deploying and maintaining tooling and processes to increase the quality and quantity of data available across the company. Enabling the development of computer vision models, complex software features, and executive-level data-driven decisions. Key tools: AWS, Dagster, dbt, Docker, GCP, GitHub Actions, Pantsbuild, PostgreSQL, Python, Terraform • Acting as Tech Lead for the data function, owning technical direction and delivery. Collaborate with Product, Software, and ML leadership to scope and prioritise data initiatives, and engage regularly with C-suite stakeholders to align on priorities and report progress against roadmaps. Regularly make trade-off decisions between cost, speed of delivery, software stability and risk. • Architected, implemented and own the company’s data platform. Deliver trusted, production-grade datasets and dashboards supporting computer vision model training, observability, performance analytics, executive reporting, and regulatory reporting. Pipelines are fast, cost-efficient, automated, and auditable. • Architected and maintain the data function’s monorepo, including CI/CD pipelines for libraries, cloud-deployed applications, infrastructure, and internal tools; enforce code reviews and leverage automation to deliver frequent, production-ready updates. Designed and implemented automated development environment configurations for various contexts, enabling fast, reliable iteration and reducing friction when building and testing new features. • Identified and remediated potential security vulnerabilities across the company codebase, including overly permissive IAM roles, hard-coded credentials, and plaintext secrets in CI/CD workflows. • Actively manage relationship with our labelling partner to maximise the quality of training data, ensure cost-effectiveness and maintain a productive feedback loop between ML Engineers and labellers

• Streamlined the model testing pipeline used by a team of five ML engineers, reducing execution time by over 80% and eliminating sources of non-determinism. This improved confidence in test results, satisfied regulatory requirements and boosted the team’s pace of delivery. • Collaborated with Operations leadership to develop tools and procedures to improve the collection and processing of computer vision training data. Reduced mental overhead and workload required of engineers, improving the reliability of expensive data collection exercises and enabling the collection of significantly larger datasets.

Delivered data analyses, dashboards, automation, tooling and machine learning projects for product teams, marketing teams and executives across the company to support data-driven decision making processes. Key tools: Airflow, ClickHouse, GitLab CI, Kafka, Pantsbuild, Python, PySpark, Tableau • Responsible for all data requirements of a product within the company. Worked with project management, development and marketing teams to create metrics for monitoring product growth, feature usage and frontend/backend performance. Dashboards and analyses regularly used to refine growth market targets, identify future features and detect bugs and performance issues. Delivered self-service growth scenario forecasting tool. • Worked with marketing team to develop metrics for monitoring the success of the company’s recent re-brand. Consolidated complex definitions into a simple Tableau dashboard to provide the team with self-service monitoring tools. Provided statistical analyses post-launch to determine the significance of observed changes. Used evidence-based arguments to challenge conclusions reached by senior executives. • Supported accounting team with automated tooling, including an income forecasting model used as the basis of annual budgets and an automated VAT calculator. Where required, implemented auditable pipelines. • Significantly improved test implementation and coverage across a range of internal tools. Improved code quality across similar tools through the application of Software Engineering principles. Implemented type checks in major shared libraries. • Developed understanding of Software Engineering principles; particularly in the effective use of CI/CD pipelines for unit tests, and integration tests.

Introduced contemporary techniques in data processing, machine learning and visualisation across the bank to improve existing processes and develop new capabilities. Successfully delivered projects in financial crime detection, demand forecasting and cybersecurity. Key tools: AWS, LightGBM, Python, PySpark, XGBoost • Led a proof-of-concept project in cybersecurity from conception through stakeholder consultation to delivery of experimental results. Extracted, cleansed and transformed data from raw text files to create a high quality dataset. Implemented academic research to develop a machine learning model capable of detecting proposed cyber threat. Communicated findings to peers and senior stakeholders. Fostered effective relationship with cybersecurity team, leading to discussion of future collaborative projects. • Jointly led development of a demand forecasting machine learning model, navigating challenging stakeholders to demonstrate a potential improvement in performance of 50% over the existing process if productionised. Presented results to peers, stakeholders and senior executives. Combined datasets from various internal and public sources. • Responsible for the line management of three colleagues. • Gained experience in the application of contemporary data science techniques, including panel data forecasting and anomaly detection, using common libraries such as LightGBM, Matplotlib, pandas, scikit-learn and statsmodels. • Developed knowledge of tools for processing large volumes of data, including AWS, Hadoop, SQL and (Py)Spark. • Produced and led presentations, workshops and visualisations for communicating complex concepts and results to stakeholders, executives and peers.

Developed navigation software for a pod to take part in the 2019 SpaceX Hyperloop competition in California. • Gained experience in applying C++ to complex practical problems • Led the experimental evaluation of various implementations • Worked collaboratively using an established version control pipeline

Developed a data driven correction to the calibration of a key component used by one of the Large Hadron Collider's major detectors. • Work implemented in active CERN research, improving accuracy of future measurements • Gained extensive experience in using Python and C++ to analyse large scale datasets • Attended a lecture programme which included classes in Machine Learning and its application to Particle Physics

Developed a data acquisition method for analyzing a prototype detector to be used in Particle Physics experiments. • Method and code used by colleagues to collect and analyze data used in published work • Accredited for work in conference proceedings • Developed programming skills in LabView and Python for data acquisition and handling

Responsible for the maintenance of several low key projects including the team’s migration to Office 365. Responsible for minutes and actions of larger meetings. Developed a range of professional skills applicable to many areas of collaborative work.

Key courses taken: Programming/electronics: Computer Modelling Data Acquisition and Handling Functional Programming Introductory Applied Machine Learning Maths: Computation and Logic Fourier Analysis and Statistics Linear Algebra and Several Variable Calculus Physics: General Relativity Quantum Mechanics Relativistic Quantum Field Theory

Selected as one of three Edinburgh students to study at Caltech for three terms. Included active research experience with the Caltech Precision Timing Laboratory - part of the CERN CMS collaboration - in characterising a Silicon Photomultiplier. Key courses taken/audited: CS011 - C workshop CS011 - C++ workshop CS156 - Machine Learning Ph020 - Computational Physics Laboratory ACM011 - Introduction to MATLAB and Mathematica