About
A modern day optimisation scientist should break down business problems like a mathematician, code like a software engineer, deploy like a data engineer, and communicate like a business analyst. I build end-to-end airline decision optimisation systems with this reliable stack.
My ownership does not stop at the mathematics. The business problem is only solved once data governance, integration, cloud deployment, CI/CD, monitoring, and the tooling's UI/UX are delivered too.
Experience
- Decision-support ownership: technical owner of a module within a production-grade decision-support software product, covering optimisation, data ingestion and post-processing across the pipeline.
- Prototyping: proving new use cases in Python before building them out as production features, weighing implementation time against runtime and technical debt.
- Cloud and CI/CD: deploying optimisation and ML modules through cloud pipelines with automated testing, logging and error handling to keep operationally critical systems reliable.
- Stakeholder engagement: working with operations teams and product managers to gather requirements, agree trade-offs and drive adoption of new tooling.
- Agile squad: contributing to code review, technical documentation and prioritisation as part of a cross-functional product squad, alongside mentoring junior team members.
- Network optimisation: metaheuristic approach to measure the cost of postcode sector to site assignments, and find optimal geographies under various demand and network scenarios. Actively used by network planning team to place new fulfilment sites and re-draw delivery areas for existing zones. Saved dozens of hours of analyst work weekly with the new approach, and enabled significant cost-per order improvements. Combined speed of low-level code approaches with metaheuristics to prioritise speed and parallelisation. Awarded: Donald Hicks Scholarship
- Sales forecasting: gradient boosted trees forecast hourly sales for three retailers across two timezones. Cut forecast error from 10% to 7% at Kroger, 5% to 2.5% at Morrisons and 6% to 2% at Ocado Retail, roughly halving again in seasonal peaks. Cloud pipelines provide fresh sales info that improve model accuracy as day proceeds. Full ownership of CI/CD, model storage, data artefact management in cloud storage. The same approach, carried into a separate road-speed model, cut budget forecast error 15% against the incumbent method.
- Rostering: CP-SAT solver for pickers and drivers, two-phase process that ensured coverage of shifts, fairness rules for workers, working time laws and the union agreement. Exports straight to payroll. Part of an initiative to increase retention at Ocado's facilities.
- Shift optimisation: combined metaheuristic search with a Monte Carlo simulation of past routes to search for optimal mixes of shift lengths at sites. 10 to 15% more drops per eight hours at some sites where the new shift mixes trialled.
- Unified cloud platform: designed a framework that puts all the data science team's optimisation tooling behind one API. Solved the problem of a 'deployment lag' when new tooling is created; if it is built to fit the API contract then new optimisation tooling can be deployed immediately. The framework integrates cloud-provider user permissions management, so data governance and data silo requirements are met.
- VRPs: helped ideate parallelisation and new algorithms to the team owning the optimisation framework which routes the tens of thousands of vans used by the Ocado Group's retail partners.
- SWE: development of heuristic architectures in Scala for a SoTA routing solver trusted by multi-national haulage companies such as Kuehne + Nagel.
- Heuristics: improved optimiser performance while decreasing run-time, presented at the VeRoLog conference Jun 2025.
- DNNs: work on PyTorch models that predict loading times for trucks depending on the content of the load, providing better bounds for the overlying optimisation model.
- Stakeholder engagement: secured new business from cash-and-carry providers at industry conferences.
- Agile work in interdisciplinary teams; taking mathematical concepts from client spec to development to deployment.
- ML & Databases: built DNN pipelines for classifying card properties and tracking stock in SQLite.
- APIs: created RESTful APIs that allowed stock to be listed on e-commerce vendors seamlessly.
- Robotics: basic card-sorting robot arm prototypes for sorting automation.
Education
Dissertation: "A Multi-Objective Evolutionary Search Strategy for Feature Selection in Machine Learning Models"
Developed a novel evolutionary algorithm approach for feature selection in ML customer churn classification for Vodafone Data Analytics. The work demonstrated significant improvements over traditional feature selection methods across multiple benchmark datasets, and is available as a pre-print.
🥈 Runner-up in the Operational Research Society's May Hicks Award
BSc Chemistry
Awards
In recognition of my work in network optimisation modelling at Ocado Logistics, I was awarded sponsorship by the OR Society to attend the 2026 IFORS Triennial Conference in Vienna.
I was awarded a runner-up award for my MSc Dissertation in Industry with Vodafone Data Analytics. Collaborating with stakeholders, I built an ML model to reduce customer churn by eight percentage points more than the incumbent method.
Awarded for services to the Department of Chemistry; I arranged surveys with students, enabling the staff committee get feedback from students and improve on NUS student satisfaction. I organised the focus groups and collated the results, which were enacted the following academic year.
Engagement
VeRoLog 2025 Jun 2025
At the University of Trento in Italy, presented research on behalf of Optrak that explores advanced search methodologies for efficiently tuning multi-phase parallel hyperheuristic optimisers with an extensive library of operators. Search approaches included Bayesian Optimisation and many metaheuristic variants.
SlidesIFORS Triennial 2026 Jul 2026
At the University of Vienna in Austria, presented the network optimisation work built at Ocado Logistics: a metaheuristic that shapes fulfilment site catchments and picks new site locations. Attendance co-funded by the Operational Research Society's Donald Hicks Scholarship.
LinkedIn
Guest lectures at several universities and colleges, promoting Operational Research as a discipline. This includes workshops arranged with the school, talks and online events. These are on my own initiative, and not affiliated with any of my employers. Schools visited: Richmond College, University of Edinburgh and Hertford College.
Leading workshops for Ocado's analysts, including sessions on Agentic AI, classical ML algorithms and advanced analytics concepts. We covered almost the entirety of the Orange Book of ML by Carl McBride Ellis as our course content. I started and carried on these sessions due to finding joy in teaching others. Consequently, analyst work has had a noticeable shift toward ML approaches and optimisation, with that mindset showing up in routine work.
While at Ocado Logistics, I built and ran an internal forecasting competition to encourage a data science approach to problem solving within analytics teams. The problem was a long-term ML forecast of sales across Ocado Retail's UK network. Participants used theory learned in ML workshops to put together a diverse range of approaches. The scoring and leaderboard updates all ran automatically via sensible CI/CD and database design. Key management obfuscated the test set from participants. Key ideas were adopted for the real long-term forecast used in planning processes.
Timeline
New Role: Senior Product Data Scientist
Sep 2026British Airways
New Role: Senior Data Scientist
Jul 2025Ocado Logistics
New Role: OR Software Developer
Sep 2023Optrak Vehicle Routing Software
Graduated: MSc with Distinction
Sep 2023University of Edinburgh
Started: MSc Operational Research
Sep 2022University of Edinburgh
New Role: Software Engineer
Oct 2020WeBuyAnyCard
Graduated: BSc Chemistry
Sep 2020University College London