Artificial Intelligence & Machine Learning Overview

Fellow presentation and intro to your career path

My name is Kai and I am originally from London, UK. I studied Maths and Physics during my undergrad, but always knew that I wanted to make the leap from science to technology. Unfortunately, I did not learn many practical software or hardware engineering skills during my degree, which mainly focused on theoretical physics and pure mathematics. I therefore decided to pursue a master's degree in Machine Learning, which provided me with a great opportunity to both build on top of my theoretical background and learn some tangible skills that I could take to the job market. My current position is working as an applied AI researcher in autonomous robotics. I deal mainly with building, developing and implementing state of the art machine learning algorithms and pipelines for robot perception.

These days, every company under the sun is trying to incorporate AI into their tech stack. Depending on your desired role, it can be very possible to find a job in the space with only a Bachelor's degree. For example, data science positions or software engineering/MLOps roles often do not require an MSc. These jobs focus more on analysing the output of ML models (data science) or building the infrastructure to build, train and deploy models (MLOps). However, if you are seriously considering a career in AI or machine learning and want to work on cutting edge algorithms and technologies, pursuing a postgraduate degree in the subject would be very beneficial. 

‍

AI/ML career options 

Here I will list a few possible career options that one can pursue in the field of Machine Learning and AI. Please note that the list is not exhaustive, and the requirements will vary depending on the position, but hopefully it will give you an idea of the different paths that are out there.

• ‍Machine Learning Engineer
  This role often involves implementing, training and analysing machine learning models for a given task. The applications can range from autonomous robotics, drug discovery, cybersecurity, augmented reality, and much more.
  ‍‍
• Data Scientist
  As a data scientist you will be mainly using off-the-shelf ML models to gain insights into how to improve your business by analysing the results of model predictions and trends.
  ‍‍
• MLOps Engineer
  MLOps engineers focus on building the infrastructure that supports ML engineers and scientists to gather, store and label data, and build, train and deploy models.
  ‍‍
• Research Scientist
  ‍As a research scientist you will contribute towards advancing the state-of-the-art in machine learning. You will work towards research paper publications in ML conferences and develop novel models and theoretical ideas. These roles will often require a PhD or equivalent experience, however, and are not entry-level positions.
  ‍‍
• Research Engineer
  Research engineers work alongside research scientists to contribute towards publications. Their responsibilities often have a lot of overlap, but where the scientists focus more on developing the theory for the project, the engineers focus on implementing the algorithms and running experiments. It is often possible to get research engineer roles with just an MSc degree, or a bachelor’s for exceptional candidates.

Main hard skills you use on daily basis in your current job

• Python programming
  I taught myself to code before and during my MSc degree. Almost all machine learning code is written in Python (some is in C++ if you are working with embedded systems or simulations).
  ‍
• Mathematics/ML theory knowledge
  This skill is imperative to be able to understand and solve machine learning based problems as an AI engineer. Whether it be reading about the latest models from research papers or choosing which model would be best suited to your problem, it is necessary to have an understanding of how the models work under the hood. It can be learned through university or online courses, such as Coursera.
  

Soft skills you use on daily basis in your current job

• Communication
  When collaborating with a team to solve a difficult engineering problem, you must be able to formulate clear and concise definitions of the problem and be able to describe your ideas and solutions at both a high level and in technical detail.
  ‍
• Analytics
  It is important to be able to determine why and how a model is failing when it is not performing as expected. This is often through extensive data analysis, plotting the results and statistics of the model and visualizing them to understand the problem better.
  ‍
• Time management skills
  ‍ML engineers must follow a timeline to complete tasks, meet deadlines, and utilize their time effectively. They can achieve it by goal setting, organization, planning and prioritization of tasks, and stress management.
  
  

Your personal path

In my personal experience, I found that looking at more than just the “known” companies was extremely helpful when finding a job. A lot of start-ups will have much more interesting opportunities than big tech companies and are often willing to compromise on experience requirements if you are a promising candidate. 

I had a very specific idea of what type of job I wanted so I only ended up applying to a handful before finding one that was a good fit, but even then, it took a few months to find the opportunity. I found my current company by Googling “cool reinforcement-learning start-ups" and looking at the careers page of each one that came up, so don’t be afraid to get creative with how you find jobs to apply for. Other than that, I found that LinkedIn offers an easy way to find new job opportunities, just make sure you check the list regularly to apply for listings as soon as they come up – I've found that you get a better response rate that way.‍‍

‍