United Kingdom

Data Product Management

Back-end

Data engineering

Architectural Engineering

Sarah Usher

Lead Architect

Problem-solving

Critical-thinking

Team player

Organisation skills

Presentation skills

Leadership

Database management

Integration Testing

Unit Testing

Python

Apache Spark

Canva

GitHub

Java

Lead Architect for BBC Data Platform and product audience telemetry

Data engineering specialist (some availability for mentorship sessions), currently consulting on a range of technology solutions for a travel startup, helping to shape innovative digital experiences for travellers and scalable, cost-effective solutions for the business.

I manage operations, communications, training & create and run workshops.

At Pleo, the Data Platform team is driving a transformative shift to a “Data as a Product” approach by implementing lineage tools, building a self-serve platform, and empowering engineering teams to take full ownership of their data. This change is fostering accountability and improving data lifecycle management across the organisation. By enabling teams to manage their own data with minimal reliance on central resources, we’re creating a more agile and scalable data culture. I’m particularly proud of how we’ve enhanced the Data Platform team’s service delivery and strengthened collaboration with Product teams, making data improvements a shared positive experience.

To scale Pleo’s data infrastructure, I launched the creation of a foundational data lake and standardised raw data storage. In addition I advocated and lead a vision for a more scalable pre-processing solution, beyond the centralised warehouse, which can be delivered iteratively, serving the most pressing data needs first.

I am also driving the adoption of a Kafka schema registry, record schema standard and schema change management process to instil a data-first mindset across the organisation. This initiative reduces errors, enhances consistency, and promotes upstream data ownership. By empowering Product Engineering teams to design and manage schemas at the source, we'll improve data reliability and reduce the burden of poor data quality downstream.

In addition to leading technical initiatives, I actively consult with Product teams on challenges like data discovery, system and application design, data modelling, and effective use of platform tools. I also mentor engineers across the business and serve on the architecture council, where I advise on high-impact, cross-functional projects.

I worked across multiple dev teams to pivot from the evolved, sprawling, hard to scale architecture to a strategic data platform/system that is scalable and fit for purpose.
I also helped the rest of the company better understand data by presenting and speaking about data concepts, so that they can feel empowered to move towards a modern data management approach, covering modern data technology and paradigms like Data Mesh.

Achievements
As IG's first Data Engineer, I played a pivotal role in the delivery of the first iteration of a new Data Platform. This enabled improved report speed and implementation of reports which were previously impossible due to data size. It also created a single source of truth which improved consistency across the application and data extracts. It involved leading a team of backend engineers and upskilling their Data Engineering knowledge so they were empowered to own and deliver the Data Platform.
I determined key domains and data abstractions to help the company create natural Data Product boundaries, which fed into the first Data Platform iteration as well as the future strategic distributed architecture design.
I ran presentations on Data Mesh Data Products and Data System History within my first few weeks, to get technical and non-technical colleagues up to speed with the state of Data Software.

SQL, dbt, Snowflake, Python, AWS, Terraform

Modernised Snyk's data processing capabilities by building Snyk's next generation, distributed data platform. Pursued a Data Mesh way of working and offer a way to achieve a minimally self-service, single-source of truth generating, data-owner managed, way of processing data. This included helping the team (architects, other engineers, managers) to upskill on modern data architecture paradigms & principles.

In addition, I tech led on a program of work to ingest package metadata from multiple ecosystems. I participated on the Testing and Python Guilds to improve these themes at Snyk. I mentored for TDD, public speaking & database themes through the Snyk mentorship program, which I founded.

Achievements
Scaled Snyks core data distribution to enable over 200% customer growth and 300% project scan growth, and beyond.
Successfully led the completion of Snyk's core data dispersion for multiple customer, custom environments.
Contributed to Snyk's first data platform and led the work to extend the platforms value by unlocking: accessible, validated data, automated system decisions, improved, reliable, scalable system architecture across the company, multiple product features and broader data & analytics capabilities. Ran multiple knowledge sharing sessions with architects & other engineers to improve their data architecture and engineering skills.
Performed technical lead & interim team lead roles.
Founded Snyk's Engineering mentorship program which expanded into more of the business and remains in place.

Data Engineering (Spark, Python, Databricks, AWS) and Backend (TypeScript, Node, Kubernetes, GCP).

Achievements: 
Designed and implemented a streaming architecture to achieve a doubling in capacity: 125k events per second, where the events reach the main reporting tool in 5mins, with the ability to rapidly scale when required.
Worked with teams across the business to get access to data not previously easily accessible.
These business teams included Data Science, BI & Analytics, Business Operations, Finance, Client Strategy & Security and Product teams.

My role:
Worked with SVP Engineering to co-ordinate and implement on Technical Data Strategy across the company
Worked with Data Scientists, Analysts and core business functions e.g. Finance & Operations.
Guided architectural design and implementation as we 
incrementally moved from batches to a more modern and innovative streaming architecture for our global data platform spanning US, UK and APAC
Implemented large scale pipeline migration, data modelling (unstructured, semi-structured & structured data), governance 	requirements and client-specific distribution jobs.
Vendor discovery, evaluation & management.
AWS Cost management using data retention policies.
Implementation of SLIs, SLOs and SLAs.
Guided-by-example code quality and disciplined practices
1 on 1 mentoring for career path and skills development for team members


In addition:
Mentored developers outside of my team as well (TDD & Java).
Ran Blogging & Speaking squad
Organiser and MC for recruitment Open Day
Created ‘ProDev’ catch-up to improve cross team relations.
Internal Tech Talks: Working with Remote Teammates, Why to Give Tech Talks, Team’s working practices (sharing how we improved delivery).
Left role due to retrenchment of whole software department.

AWS (Kinesis, EC2, EMR, Lambda), Java, Python, Spark (on EMR)

Achievements:
Successfully lead a data integration and quality project for the Data Science team.
Completed further integrations, providing data to the business in a way previously done manually. 
Created the data lake with a standardised process to change data for consistent data across the business.
Optimised the pipeline for increased data capacity.

The Reporting and Data Team is responsible for: 
Architecting data pipelines & data stores.
Ingesting, processing and storing all data in evolvable formats.
Maintaining and increasing data pipeline speed and capacity.
Providing appropriate and timely access to data for Data Scientists, Analysts, Business Operations and Product Teams.
Ensuring strict cost management as we have a very large data-set.

My role involved:
Ensuring good technical behaviour (good code, tests, thought-out architecture).
Tool discovery.
Lead on a stream of work (as opposed to Tech Lead which is technical oversight of all streams of work).

AWS (Kinesis, EC2, EMR, Lambda), Java, Python, Spark (on EMR)

Improved an iOS Swift employee application – cleaner code, better structure, correcting the storage strategies, a 'continuous-as-possible' delivery pipeline (automated up to Apple approval stage).
Created a knowledge base (client-demoa-ble application and developer wiki) from which other developers could learn more about building and deploying microservice applications on the Azure platform including Docker, Kubernetes and Azure API management. 
Improved code quality of a Train machine-learning application.
Organised and ran an Introduction to Swift workshop for the Ladies of Code meetup group.

.Net developer on the Azure cloud, multi-tenant application - HyperLaw, including technical client engagements. 
•	Improved highly-coupled codebase by refactoring functionality into dedicated, single-responsibility services, introducing dependency injection and an ORM.
•	Replaced custom authentication with standard ASP NET Identity.

ASP.NET MVC

Having completed coursework part-time, I opted to complete my research report full-time.

The report explored the use of Maximally Stable Extremal Regions - an image processing algorithm - together with supervised machine learning algorithms, to remove image noise which can cause bad quality OCR output. The technique was tested by comparing the OCR output from the images before and after processing. The results show that, once the right data is selected, it is possible to improve the OCR output quality.

Research Report name: Using customised image processing for noise reduction to extract data from early 20th century African newspapers. 

MatLab, R.

Completed an Android prototype for mobile doctors to capture data during their rounds at hospitals. The application was a data and photo capturing form. The details were emailed to the doctor's secretary from the application.

WCF Services sub-ledger application for about 3 years.
Our team embraced the Sprint and then Kanban methodologies and became one of the most highly-functioning teams in the department. This was a combination of committing to a process and continuing to improve on it, making good strategic decisions and sticking to our guns about our way of working. We built up high trust with our users and Product Owner because of our improving delivery abilities. This was despite being without a Team Lead for a few months! I played a good part in changing and maintaining our team's processes as the bank started to embrace a more Agile way of working. 

I'm also interested in retaining happy teams and part of my work here was heading up our social committee to organise creative boardroom-bound, video-conference functions for our departments global teams. These were very successful and made a great impact on team collaboration and general job appeal - they were good fun!

As I had built up a good working knowledge of the bank and had good relationships across departments I often consulted and assisted on other projects.

C#, MS-SQL, Silverlight, WCF

Sarah Usher🇬🇧