An insight into Kainos' AI Practice

I'm Sapphire Duffy, a recent graduate working in the AI Practice at Kainos. I knew I needed to be somewhere exciting, fast paced and challenging with problems that I could solve! Artificial Intelligence (AI) certainly delivers that for me, it is extremely rewarding. I saw an advertisement for Kainos AI Camp during university and I was interested in the big hype around AI and eager to learn more. I was delighted to get the opportunity to be a participant for AI Camp in 2018. I instantly fell in love with AI and the idea of using AI solutions to solve business problems. Soon after I joined Kainos, I moved into the AI Practice.

When we hear Artificial Intelligence what usually comes to mind is robots taking over the world, that's certainly what I first thought! I now know that AI is nothing like that and is so much more.

So, what is AI? 

Artificial Intelligence is an area of computer science ultimately working to make computers perform human tasks. Different categories of AI include Robotics, Speech and Text, Automation and Machine Learning. Machine Learning is a core part of AI and is based on the idea that systems can learn from data, identify patterns and make decisions with minimal human intervention. It's essentially the science of getting a computer to act without programming. There are three types of machine learning algorithms; Supervised, Unsupervised and Reinforcement Learning. AI has made its way into a wide range of areas including Healthcare, Business, Education and Finance - AI is changing the world. 

AI Practice Background 

The AI practice was formed within Kainos to build a capability in delivering production AI services to solve business problems for both commercial and public sector customers. Kainos has moved beyond the experimentation phase with AI, deploying AI solutions more broadly and realising benefits across our customers' business. AI solutions start with business problems, not datasets. We then set out to prove the value with a small dataset then proceed to discovery, alpha and beta service. An example of one of our successes back in 2018, was our work with DVSA to create the anti-fraud risk rating service for MOT garages using Machine Learning. The practice provides leading edge AI services on fraud, risk and automation. The AI practitioners have been recognised for their work with DVSA as an exemplar of production AI by both Government Digital Services and the office of AI. This project also won the AI Project of the Year award at the UKIT awards. 

Team Setup 

Our AI Practice is made up of AI Engineers, Data Scientists, Technical Architect and Practice Leads that are spread across four different locations; Belfast, Gdansk, Birmingham and London. Our team are responsible for developing high quality solutions which integrate AI and ML technologies into production systems, solving complex business problems and generating efficiencies and substantial savings for our customers.  

AI Engineers and Data Scientists work collaboratively within a broad multidisciplinary and are responsible for building and testing software applications that integrate AI services and include data science models. They apply mathematical principles to gain insight from data and have exposure to major machine learning frameworks, languages and NLP frameworks. They apply common design principles and patterns. They clean, filter and re-factor complex data from different sources and have experience in working with relational databases. 

We have a diverse team with a mixture of different skills and experiences which brings broader ideas and different perspectives. We have a great team culture and personally when I joined the team from the very start, I felt valued, trusted and equal to everyone else. 

Training and Upskilling  

Kainos are committed to training and upskilling their people and this remains important to our continual professional development. Within this notion we have technical spikes that allow us to continually develop our technical capabilities and share resulting findings/learnings with other team members. Everyday you learn something new which is exciting and rewarding. We have a strategic training plan in place for all levels within the practice. 

AI Club is an internal initiative that takes place at four of our key sites and gives all Kainos employees the opportunity to learn more about AI. 

If you want to find out more about where to get started in AI, a second blog in this series will be published soon! 

So, what is it like day-to-day in the AI Practice?  

Below, I have asked a few colleagues what their day-to-day job entails.  

Claire Houston, Senior AI Engineer 

I'm a Senior AI engineer which I really enjoy and usually leaves my desk covered in to-do lists. I usually program in python, with deployments on AWS and Azure. I've used machine learning algorithms, natural language processing and deep learning in my role- all depending on what's needed for the project. My role is often customer facing, working on discoveries with projects or demoing finished products. I'm currently working on a number of potential chatbot projects and demoing our POC for a financial entity extraction tool for Funds Axis, an investment management company.  

Oliver Wilson, Technical Architect 

As an architect, I am responsible for technical delivery of specific projects within the AI Practice, but my raison d'??tre is to bring knowledge and experience from "normal" software engineering into the department. A good day will involve interacting with customers, discerning the problems they face, and designing cost-effective solutions that will get Kainos AI into production. 

It's a privilege to work in the AI Practice, as the department is filled with experts in a field I've not previously been exposed to.  There's tremendous growth potential in this space, and an opportunity for us to be leaders in a new market segment. Compared to my previous appointments, day-to-day I do a lot less fine-grained implementation, and a lot more hunting for patterns and best practice in a space that's still very green. 

Claire Davidson, Senior Data Scientist  

Working as a Senior Data Scientist for Kainos, my day to day work is quite diverse. Some days it involves in-depth analysis and building machine learning models, as well as attending important stakeholder meetings with our customers. 

Project work often involves a combination of data mining, data visualisation, machine learning and text analysis using tools such as Python, PowerBI and SQL. The key aim is to unlock insights from our customer's data and understand what possible advanced analytics techniques could be applied to improve business processes. Collaboration with our customers and other team members is also a vital part of the job to ensure the resulting service meets our customer's requirements.

 I am currently working on a document comparison tool which applies text analysis techniques to text extracted from PDFs using AWS Textract. I have also worked with DVSA to build a machine learning tool which assists in the detection of poor standards and fraudulent activity within MOT garages across Great Britain, resulting in a positive impact on road safety. 

Projects  

So, you may be wondering what kind of projects we do? Kainos has delivered a wide range of AI projects including:  

DVSA ?�� DVSA conduct 42,000 tests per year and their biggest challenges are to maintain standards and eliminate fraud. This project won UKIT award in 2018 for the best AI project of the year. To find out more about this project, check here.  

Smart Chatbot ?�� For this project, we used Natural Language processing and our team have created a chat-bot that helps with answering questions regarding WorkSmart sales, based off a dataset of previously asked questions and responses.  

Clinical Autocoding - we worked with RBCH to investigate the potential to automate manual clinical coding. It is estimated that ??30,000,000 is spent by the NHS each year on manual clinical coding and ??400,000 per annum cost for RBCH alone. >90% accuracy achieved which massively exceeds human level accuracy. 

Conclusion

Now you see why the AI Practice really suited me? It's exciting, fast paced, challenging with many problems to be solved and it is extremely rewarding!

Have a look in the careers section on the Kainos website, we are actively looking for passionate, enthusiastic, hard-working Lead AI Engineers and Data Scientists.  

You can also come and meet some of the team at AI Con, which is taking place on Thursday 28^th November in the Europa hotel, Belfast. The conference is the first of its kind in Northern Ireland and is bringing together world-leading technology professionals and business leaders to discuss how AI is changing the world.