Kainos Final Year Projects — Showcasing Innovation
Supporting talent development and providing educational programmes is at the core of Kainos culture and values. Through various schemes, we aim to provide a platform in which education, industry and innovation can come together, to deliver valuable experiences for those starting their career journey.
As the Applied Innovation team, it’s important we are helping to cultivate an environment that encourages and develops innovative thinking to deliver unique solutions to industry problems; that help keep Kainos looking forward. The Kainos Final Year Project program is one of the ways we help support this. Through developing these ideas we are promoting the idea generartion process our team takes and knowledge sharing around the potential impact and value these new technologies could have.
So what is the Final Year Project Scheme?
Each year we provide a range of project ideas for final year students, in collaboration with Queen’s University Belfast, that are in new technologies or future areas of interest. By providing these projects, we aim to make sure that the students are working on unique problems; that also create a challenging and engaging experience for the students. It is common that these projects align closely the Innovation teams goals, such as new areas of technology we are interested in or that would be valuable for our customers, which drives our dedication to commit time and mentors to the projects.
Throughout the course of the year, each student is individually mentored by an employee from an area of the business; this mentor will help guide them through their project. The mentoring process can take can various forms, such as technical help or demonstration and presentation skills. By creating mentoring opportunities for employees within Kainos, we are able to promote the development of leadership and management skills that are invaluable for their careers. As a company, this opportunity to work closely education institutions and their students, is a brilliant way to cultivate and support the talent that is coming into the industry.
The Students, the Projects and the Success
This year saw the completion of 5 projects, all of which have a focus on AI related topics. These span a wide range of applications from tools that can increase the size of an AI’s traning dataset, to the Auto-redcation of Personally Identifiable Information.
Chloe McAteer: Sensitive data filter for legal systems
Increasingly, consumers are becoming more aware about their personal data. With the introduction of new regulations such as GDPR, companies have their reputation and financial standing on the line if customer data is not held securely. With AI systems needing context rich data in order to use common features to make decisions; these regulations often add complexity when it comes to acquiring data that can be used to train AI models.
Chloe’s solution can ingest a series of documents and auto-redact the personal details within each document. This tool has significant potential in both the private and public sectors by rendering text data in a GDPR compliant state, which can then be used within further AI solutions. With the redaction of information such as names, organisations and currency, this tool can drive data anonymization at scale, which ultimately leads to better data security and greater value for end users.
This project was also awarded the IET Megaw Memorial Award, for the best paper submission and 10 minute presentation on a final year project.
Chloe Mullan: Intelligent object removal and replacement
Home hunting is a challenge we all find ourselves facing at some point in life, but it all boils down to a single, critical question. Can I see myself living here? We naturally try to visualise ourselves in the space but when viewing a home its hard to look past the incumbent furniture, decoration and style to gain an appreciation for a space.
How can AI help? Chloe Mullan’s project involves leveraging a machine learning algorithm known as a Generative Adversarial Network to automatically remove furniture from rooms. The main goal is to improve the viewing experience for home buyers and to make self projection into the space possible. Being able to view the space as a blank canvas and visualise its potential, increases the chance of selling and therefore is a benefit for both sellers and real estate agents.
Christopher Burns: AI Model operations and management platform
AI is rapidly maturing and becoming more mainstream, so much so that companies with a non-technical background are wondering how they can leverage AI solutions. They are hoping to gain insights into their data and customers, to drive business or create value for their customers. A common issue around the use of AI is making it accessible to non-technical roles and the lack of support for model deployment; especially in circumstances where a model should be updated over time.
Christophers’ project focuses on the operational side of managing machine learning models, through the use of a web based application. This application acts as a fully wrapped platform for model training, testing and deployment; making it a valuable solution for non-technical users. This platform has significant potential for how we modify models, in relation to retraining on new data and reducing the pain of version control during deployment.
Gareth Dunne: Vision first approach to software testing
Workday is one of the largest human resource management tools in the world. Kainos have developed and support, an automated testing tool for Workday known as Smart. To maintain the Smart product, it frequently needs to adapt to changes to the Workday platform; these can range from visual modifications to significant alterations in the structure of the web application. This continuous delivery involves development, testing and redeployment of the Smart tool, which is expensive for the internal Smart teams within Kainos.
Gareth’s solution is a web application that makes this process far more efficient, through leveraging AI and complex Computer Vision techniques. This platform can be used to monitor changes to Workday, using a vision based understanding of the webpages and then modify the Smart testing tool as required. This project has transformed a heavily manual, time consuming process into an automated, monitoring focused workflow, freeing the Smart development teams to focus on higher value features to the tool.
Tiarnan Cooney: Simulation to generate artificial datasets for automation
Datasets are one of the most important parts to creating fully functional AI models, but increasingly the availability of data is becoming more apparent. Collecting data is time consuming, computationally expensive and often it is difficult to ensure that we are gathering not just quantities of data but high quality data.
Tiarnan’s final year project focuses on how we can avail of simulation techniques in collaboration with AI models to build data generation and recognition systems. Using Unity3D and 3D assets, this solution has enabled the generation of large synthetic datasets by simulating real world objects and capturing these as images from various orientations. These images have then been used with Javascript based AI tools to develop an AI that can identify these objects in a vision based manner.
Lights, Camera….Showcase
As part of the final year project scheme, we aim to provide the students with as much interaction with the business as possible. At the end of their projects Kainos hosts an event to showcase and celebrate the hard work of the students. We use this time as a way to knowledge share and promote innovative solutions to various areas of the company, with the hope to inspire and inform, while also giving the students a platform to promote their work.
By involving stakeholders from across the business, the students are able to demonstrate their projects and help the company understand the impact that these solutions could have on our customers. This showcase event is also the starting platform to spark ideas for following years projects.
Belfast is a hub for future talent and it is a brilliant opportunity for Kainos to support and mentor young professionals through their university careers. All of these projects have provided valuable insights for Kainos into applications and value of AI solutions; that can then progress on to develop business value either internally or for our customers.
We of course have to thank the students for their hard work and dedication over the year and congratulate them on their outstanding First class results; hopefully these projects have created value for the students as they continue into graduate employment. Lastly, this program would not be possible without the support and invaluable time from our mentors, Liam Ferris and Jordan McDonald.
Want to get involved?
If you are a student going into final year at Queen’s University Belfast and are interested in the opportunities Kainos has for Final Year Projects, please get in contact your University module supervisior. We would also love to hear from you if you are interested in hearing more about this years projects, or if you would like to hear about our mentoring opportunities. If you are a university looking to partner with us you can contact us through our website.
Keep an eye out for next years ideas, we’re looking forward to another group of students and exciting projects!