SPAICE Tech
London
The placement allows me to work closely with a broad range of data and signals. From camera images, LiDAR pointclouds to RADAR signals, my job is to develop algorithms that can synthesise and understand this data to be used to help navigate and understand challenging environments. The job lets me work on the cutting edge of perception and autonomy, to build software for vehicles from helicopters all the way to satellites.
As a Computer Vision Engineer, the role demands strong mathematical, coding and problem solving skills; it is a truly encompassing role. This placement has let me work on exciting tech stacks that run on applications that will be run over 30,000km from the earth. Deploying and testing these solutions we develop is a rewarding challenge. My contributions are a mix of working with AI and traditional software solutions. Developing novel machine learning pipelines on sensor data, developing probabilistic deep learning models that can be deployed on edge devices and deploying cloud applications tools for client facing and internal needs are some of the high level contributions. I have pushed client projects forward and led many new ones. Being at a smaller company gives me the challenge and experience with working and leading projects and having greater direct responsibility on the end product.
The role required me to leverage my mathematical skills and techniques I learnt at university. Applying theory to practical real world problems deepened my understanding and intuition behind core statistical and mathematical concepts. I learnt a lot of new machine learning approaches to handling tricky data and building robust signal processing techniques. I faced challenges generalising my pipelines to broad scenarios, which taught me to build resilient code and work iteratively to improve solutions.
The role required me to keep up to date with the newest industrial and academic solutions available on the market, teaching me how to improve my comprehension of difficult concepts, especially in tricky research papers. It also taught me how to communicate my ideas efficiently to both technical and non-technical members. One thing I learnt is how to be able to accurately evaluate my own level of understanding of concepts - to not convince myself I know something until I can communicate and think deeply about problems surrounding the topic.