As an AI Software Engineer, I have seen firsthand how rapidly AI development has evolved in my profession. Looking back to around 10 years ago, utilising the most advanced AI technologies was the domain of academia and a few large corporations. While the majority of the companies relied on simplest AI methods based on the cost-benefit tradeoff. The invention of Large Language Models (LLMs) and development of new tools democratised the use of most sophisticated AI techniques for users worldwide. This shift led to the remarkable AI advancements we've witnessed over the last year.
My path into AI and machine learning was somewhat serendipitous. Pursuing degrees in Applied Mathematics and Physics, I was drawn by my love for maths and programming. However, I soon realised that developing numerical methods to simulate physical processes was not my calling. I wanted to find an area where maths and programming overlapped and interacted, but without connections and mentors working in similar fields I was not aware of the breadth of opportunities available to me.
My real passion emerged during a class where we explored machine learning applications in a related subject. This exposure opened my eyes to the vast potential of machine learning and AI across various disciplines. Consequently, I completed my PhD in Language Technologies at Carnegie Mellon University in 2013, with a focus on natural language processing. Since then, the advent of large language models has dramatically reshaped the field. I would not be surprised if large parts of the curriculum are entirely different today.
In my professional life, I've primarily focused on recommendation systems. My seven-year tenure at LinkedIn and the last 18 months at Handshake have been particularly influential. At LinkedIn Learning, I was part of the foundational team that developed systems to connect users with educational content to help them advance their career. The mission of a company or a project is of extreme importance to me. When I first joined LinkedIn I was offered a number of different projects to work on and I did not hesitate for a second when LinkedIn Learning was on the table. The project's potential for positive impact was a major draw for me, much like my decision to join Handshake.
At Handshake, I've been involved in projects like automating employers’ outreach to students, enhancing the new student feed with personalised content relevant to individual needs and interests, as well as the initial exploration work when we began discussing the development of Coco, an AI-assisted career guide. I believe these tools can have an enormous impact on students like me who were not aware of all the options available to them throughout their studies. The challenge lies in creating models that accommodate individual nuances and rare scenarios, which can be addressed through specialised datasets and precise user prompts. Users either need to be trained or the model needs to be designed with that unique use case in mind.
The transformative potential of AI in bridging the gap between students and employers is an aspect I find particularly intriguing. AI tools, with their advanced capabilities, offer a streamlined path for students to enhance their marketability. They serve as digital mentors, guiding students in crafting compelling resumes, profiles, and applications, tailored to resonate with prospective employers. This guidance is not just about formatting or phrasing; it's about understanding the nuances of what makes a candidate stand out in a competitive job market.
Conversely, for employers, AI paves the way for creating job descriptions and outreach messages that are not only more appealing but also more inclusive. This is a significant leap towards fostering diverse and dynamic workplaces. But beyond these immediate applications, there's an emerging horizon that captures my enthusiasm: the realm of personalised educational materials.
Imagine an educational landscape where curriculum, lecture notes, and assignments are not static, one-size-fits-all materials, but dynamic, tailored resources shaped by generative AI. This concept moves beyond the traditional paradigms of education, offering materials that adapt to individual learning styles and current industry trends. The inception of this idea is already taking root in the initiatives of pioneering edtech startups, such as EduAide and Nolej. These ventures are not just experimenting with AI's capabilities; they are trailblazing a path towards a future where education is more personalised, adaptive, and aligned with the evolving demands of the job market.
My day-to-day activities have evolved significantly. Nowadays, staying abreast of the latest research and development frameworks in AI is crucial. The informational resources I rely on have also changed, moving from API documentation, to community sites like Stack Overflow, and now to integrated co-pilot tools. Generative AI tools also enable me to automate routine tasks like documentation and emails, allowing me to focus on more critical aspects of my work.
Despite these advancements, it's important to recognise the limitations of AI tools. Precision in input is key to obtaining useful outputs. I always review, modify, and verify AI-generated content before its final use. I view these tools as partners in creativity, providing a foundation for my ideas that I can refine with my expertise and knowledge.
As we peer into the next decade, I anticipate an era where AI reshapes our professional landscapes. Requiring us all to be lifelong learners and making many existing professions more productive. Those currently in education often adopt technologies faster, giving many students entering the workforce today a distinct advantage.
Sign up to our monthly Early Talent and AI newsletter to keep up with the latest news around AI and the future of work from trusted sources as well generative AI tools and resources that help you do more with less.
