Recently, there has been a significant shift in the area of large language models (LLMs) and AI-enabled applications. With the advent of foundation models like GPT and Llama, and most recently, the Segment Anything Model (SAM), the need for in-house training of AI models has diminished for most organizations. 

As a result, the skill set required to build AI applications and tools has evolved, moving away from academic and research-oriented expertise and towards a more practical and application-focused approach. This shift has led to the emergence of new roles in the AI and software engineering domains, such as the Applied AI Engineer. This blog post will discuss how these changes impact the software development industry and explore the skills now in high demand.

The Changing Paradigm of AI Development

The capabilities of AI and machine learning models have grown exponentially, with modern foundation models becoming increasingly accessible to organizations. Organizations can fine-tune and adapt these models to various specific applications through techniques like prompt engineering, removing the need for most organizations to train large AI models in-house.

While smaller companies may eventually be able to train their own large AI models, the current focus has shifted towards leveraging the power of existing foundation models to build functional, practical applications. Due to this, the capability to deliver AI software solutions is more critical than ever. Consequently, the skills required to construct AI-enabled applications have also changed, with a greater emphasis on software engineering expertise rather than theoretical AI research. 

The Emergence of Applied AI Engineers

As the industry evolves, so does the demand for new roles in software engineering. One such emerging role is the Applied AI Engineer. These individuals possess strong software engineering skills and a solid understanding of AI and machine learning concepts. Their primary focus is to design, develop, and implement practical applications built on top of foundation models, integrating AI capabilities into real-world use cases.

These engineers must be skilled in various software engineering techniques similar to those of a backend software engineer — understanding cloud computing, cloud infrastructure, and general cloud-native software development concepts is beneficial in addition to general software engineering skills.

To be able to design and implement AI software solutions that integrate with state-of-the-art cloud software, applied AI engineer skills include but are not limited to the following areas and technologies:

• Understanding of AI and machine learning concepts

• Model fine-tuning and transfer learning

• Prompt engineering and model adaptation

• Vector databases, such as Qdrant and Pinecone

• Open-source software libraries and tools

  • For example, LlamaIndex, LangChain and Guardrails

  • Python AI and ML ecosystem, in general

• Model evaluation and performance measurement

• Greenfield software development and integration to new and existing solutions

Based on this, a new area of software engineering specialization emerges: applied AI. The Applied AI Engineer role fills the gap between AI Researchers and Backend Software Engineers.

Applied AI engineer is a specialized role that bridges the gap between AI Researcher / Data Scientist and Software Engineer.

AI Researcher & Data Scientist

AI Researchers and Data Scientists primarily focus on the theoretical and experimental aspects of AI and machine learning. They develop new algorithms, improve existing models, and explore the underlying principles of AI. Their skillset is heavily research-oriented, and they are often involved in publishing papers, attending conferences, and advancing the state-of-the-art in AI. While their work is essential for pushing the boundaries of what AI can achieve, it is often less focused on direct practical applications.

This role is crucial for organizations that develop foundational models. However, it will likely become less critical for most small—to mid-size organizations in the future as the availability of foundational models increases. Data Scientists will focus more in the practical aspects of AI in the future.

Applied AI Engineer

Applied AI Engineers bridge the gap between AI research and practical applications. They have a strong foundation in software engineering and AI concepts, allowing them to leverage and apply the latest research findings to real-world use cases. Their main goal is developing, fine-tuning, and adapting AI models to create practical applications for businesses and consumers. While they may not be as deeply involved in theoretical AI research as AI Researchers, they play a crucial role in bringing AI capabilities to life in functional applications.

This role is vital when utilizing and integrating the latest foundation models into new and existing applications. This role is needed in many, if not all, organizations that apply AI to real-life applications.

Backend Software Engineer

Backend Software Engineers focus on developing, deploying, and maintaining cloud applications and systems. Their expertise lies in designing and implementing the underlying software infrastructure that supports web applications, mobile apps, and other digital services. While they may not be directly involved in AI development, their work is essential for ensuring the smooth operation and integration of AI-powered applications within larger software ecosystems.

This role has been and will continue to be essential for any organization that develops software solutions.

The Future of AI and the Software Engineering Workforce

The recent AI developments have brought new challenges and opportunities for the software engineering workforce. As the focus is moving from training large AI models in-house to leveraging pre-trained foundation models, the need for professionals with both engineering and AI expertise will only continue to grow. Applied AI Engineers and other emerging roles are at the forefront of this transformation, bridging the gap between the latest AI capabilities and practical, real-world applications.

In the short term, most companies cannot fill the skill gap as fast as the AI world is moving. Organizations must seek external help to leverage AI and foundation models in their operations to overcome this challenge.

At Softlandia, we will soon deliver an applied AI solution for a customer. This project, built on top of a foundation model, showcases Softlandia's ability to create practical and efficient applications that harness the power of AI and large language models. Stay tuned for a detailed customer case post that will be published soon! In the meantime, check out our tips about NLP solution building and sign up for our next data science infrastructure meetup focusing on LLM’s. Check out also the recording and recap of the previous meetup.

Don't hesitate to contact us, if your organization needs assistance with applying AI or integrating it into your existing systems and workflows. Our experienced consultants are ready to help you enable AI capabilities in your applications, ensuring your business stays ahead in this rapidly evolving field. Contact us today to discuss how we can support your AI journey and unlock the full potential of AI for your organization!

UPDATE 21.08.2023: In case you are looking more for an enterprise-ready solution, we've got you covered as well! Beta launch of our product YOKOT.AI Generative Enterprise Chat & AI is now done and is available for demo.