Academic Foundations in Security and Crime Science
I completed a Bachelor of Science in Security and Crime Science at University College London. This programme blended various disciplines, including criminology, social research, psychology, and mixed methods research. Additionally, it incorporated coding aspects such as statistics and machine learning, and I opted for modules in web development and simulation. A notable component of my degree was a six-month work placement where I served as a Research Intern at the Mayor’s Office for Policing and Crime, specifically in their Evaluation unit.
Discovering Operational Research
My introduction to the Government Operational Research Service (GORS) occurred after I had begun my career in the Civil Service. It was during this time that I learned about Operational Research (OR), which made me realise how closely related it was to my academic background, particularly in relation to evaluating crime prevention strategies.
Motivations for Joining GORS
A significant portion of my studies revolved around assessing outcomes following various interventions. This focus fostered a desire to engage in projects that generate meaningful real-world impacts. With a strong analytical mindset and a passion for coding developed during my university years, I felt a natural inclination towards OR. In hindsight, it was the ideal career path for my interests and skills.
Moreover, I had a mentor from the Ministry of Justice (MOJ) whom I met through a charitable organisation during my university years. Although he was not directly associated with GORS, his insights into working within the civil service and the MOJ were invaluable in guiding my career choices.
Current Position and Responsibilities
Presently, I serve as an Associate Data Science Product Manager, where I oversee product vision, ethical considerations, stakeholder management, risk assessment, and documentation for Data Science initiatives in Probation Data Science. My current focus lies in projects that involve Large Language Models. My technical background has proven advantageous, enabling me to grasp model limitations and convey technical concepts to non-technical stakeholders effectively.
Prior to my current role in product management, I worked as a Senior Data Scientist at the MOJ. In that capacity, I engaged in various coding projects related to large-scale data linking within the criminal justice system, fines enforcement, estimating reconviction rates, extracting insights from employee data, and developing a labour market dashboard. Each project presented new learning opportunities, and over time, I have been able to recognise my growth in both technical and interpersonal skills.
Significant Accomplishments
Two specific projects stand out as highlights of my career thus far.
The first is the launch of our Contact Log Semantic Search. Probation staff generate millions of reports on offenders each year, known as contact logs. These reports are often unstructured and vary significantly in format, making it challenging for staff to efficiently search for specific information. This can hinder risk assessments and preparation for offender engagement.
To address this, we trained a large language model on extensive text data from multiple sources to ascertain the relationships between various words, phrases, and sentences. When utilised, the search tool analyses each entry in an individual’s contact log to identify relevant matches based on semantic similarity. This innovative approach significantly enhances the relevance of search results and reduces the time staff spend searching for information, thereby enabling them to concentrate on engaging with individuals on probation and achieving optimal rehabilitation outcomes.
Enhancing Data Linkage with Innovative Solutions
The second noteworthy project involved advancing our efforts in person data linkage across the criminal justice system. The administrative data within this sector is often inconsistent, with individuals frequently lacking a unified identification number or acquiring new identifiers with each interaction. This inconsistency complicates efforts to accurately estimate statistics needed to evaluate interventions and their interactions with other governmental services, especially when individuals change names or addresses.
Fortunately, our exceptional data linkage team developed Splink, a highly regarded open-source package with over 10 million global downloads. My role involved enhancing the outputs of an existing Splink person linkage by incorporating an additional data source, thereby improving our data accuracy and reliability.
Guidance for Future GORS Applicants
To join GORS, a fundamental requirement is that at least 50% of your degree must be numerate in nature. However, beyond this, curiosity is essential. I always advise those interested in a role to reach out to individuals currently in the field for insights; this often leads to valuable mentorship opportunities.
Ultimately, the specific coding language you know is less important than your analytical reasoning abilities. GORS values analytical aptitude—your capacity to logically evaluate which analytical techniques are appropriate. After conducting your research, being able to convey your findings clearly, whether verbally or through written communication, is crucial.
Success in this field comes down to practice. As long as you remain curious, disciplined, and a touch creative, you will thrive. Many university students fret over lacking internship experience; however, dedicating that time to practical coding projects can be far more beneficial for your CV and learning journey. The abundance of open-source data and the possibility of generating synthetic data provide ample opportunities for hands-on experience.