As AI proliferates, a new breed of engineers is emerging: AI Engineers. Unlike traditional software engineers, AI Engineers specialize in integrating, deploying, and optimizing AI models. They navigate complex ecosystems involving LLMs, APIs, and traditional software to develop apps.

Types of AI Engineers

AI Engineers can generally be categorized into three main types based on their focus areas and contributions:

1. Frontier AI Engineers

Example: Member of Technical Staff at OpenAI

These engineers work at the cutting edge of AI research, typically employed by organizations like OpenAI or Anthropic. They often hold titles like "Member of Technical Staff" and are responsible for designing and training large-scale models. Their work involves:

• Developing foundational AI models (e.g., GPT, Claude, Gemini)
• Training deep learning networks using TensorFlow and PyTorch
• Benchmarking and optimizing model architectures
• Managing vast datasets and computing infrastructure

Frontier AI Engineers operate at the 0 → 1 stage of AI development, requiring extensive resources, research capabilities, and financial investment—often in the millions of dollars—to build net-new AI models.

2. Infrastructure AI Engineers

Example: SWE at LangChain

These engineers specialize in deploying and scaling AI models, ensuring that foundational AI research is accessible via APIs. Companies like Anon.com, LangChain, and BaseTen fall into this category. Their primary responsibilities include:

• API development and model hosting
• Optimizing inference for speed and cost-effectiveness
• Ensuring distributed and scalable AI deployments
• Creating SDKs and tools for developers

While they don’t create new AI models from scratch, Infrastructure AI Engineers ensure that cutting-edge AI research reaches millions of developers worldwide in an efficient and usable manner.

3. Applied AI Engineers

Example: Founding Engineer at SellScale

Applied AI Engineers focus on integrating AI models into real-world applications. They leverage APIs and pre-trained models to build functional products, such as:

• AI-powered sales outreach tools (e.g., SellScale)
• AI-driven software agents (e.g., Devin, Harvey)
• Custom AI applications for various industries

These engineers work at the application layer, optimizing AI workflows, refining prompts, and ensuring that AI interacts effectively with users. Their primary focus is on making AI useful and impactful in practical scenarios.

The Core Skills of an AI Engineers in the Application Layer

Previous Key Skills:

• Logic, systems, and scalability
• Basic UI/UX
• CS fundamentals

New Key Skills:

• Liberal arts
• Prompting techniques
• Self-evaluation

What skills are new?

While the field of AI Engineering is evolving rapidly, several foundational skills are essential for success—especially in the application layer.

1. Liberal arts

• AI is fundamentally about processing and understanding language.
• The ability to craft effective prompts and interpret AI outputs is a core skill.
• Engineers with strong linguistic intuition will have an advantage in fine-tuning AI behavior.

2. APIs & JSON Proficiency

• APIs are the backbone of AI applications.
• Engineers need to understand how to integrate AI via API calls, interpret documentation, and structure JSON responses.
• The ability to modify and manipulate JSON payloads creatively is a critical skill.

3. Self-Critique

• Unlike traditional engineering, where user feedback drives iteration, AI Engineers must assess and refine model outputs independently.
• AI degradation is a major concern—feeding AI-generated content back into AI models can lead to quality loss over time.
• Vigilance in evaluating and optimizing LLM outputs is essential.

What skills are still important?

Traditional CS Fundamentals

• A strong computer science background remains a prerequisite.
• Over-reliance on tools like Cursor for autocomplete will limit deep technical understanding.
• Core skills in data structures, algorithms, and software engineering are still critical.

Conclusion

Most engineers are SaaS or application based today.

The AI Engineer will continue to proliferate. You can use these skills to grow into it.