At Gun.io, we’ve witnessed numerous companies who are looking for an AI/ML engineer, when a capable Backend developer would do. The line between AI/ML and backend development is increasingly blurred, creating a particularly challenging landscape for hiring managers and CTOs. We often find that backend developers are what companies are actually looking for, even when they think they need AI/ML expertise.
Taking a Step Back: The Evolution of AI Roles
It’s crucial to place the role of an AI engineer in the context of rapidly evolving technology trends and shifting jargon. Just two years ago, “AI Engineer” wasn’t a common job title. Instead, we had Data Engineers, Data Scientists, ML Engineers (or Applied Machine Learning Engineers), Backend Engineers, and MLOps Engineers. In fact, the title of this article would have been “Do you need a Data Scientist or a Backend Developer (or Data Engineer)?” back then.
Historically, the answer was often that companies needed a Data Engineer first, and then maybe a Data Scientist. This was because most organizations grappled with messy data and ill-defined goals. The bulk of the work involved cleaning, preparing, storing, and extracting data — all data engineering tasks. Model development and tuning typically constituted only about 20% of the work in a good project, or even 0% if the data was poor or the problem wasn’t actually a data science problem.
Fast forward to today, and we’re in the era of Large Language Models (LLMs). Now, companies almost always need a Backend Engineer (potentially with some DevOps skills) and might add a Data Engineer if they’re implementing Retrieval-Augmented Generation (RAG) — a workflow combining LLMs with a company’s internal or customer data.
It’s worth noting that “AI engineer” means different things to different people. To technical folks, it often translates to “data scientist,” while to founders, it might mean “someone who can enable me to make money using this GenAI technology.”
Given the rapid pace of change in AI, particularly with LLMs, no one has more than two years of experience in this specific field. The most valuable skills currently involve the ability to read and implement documentation from frameworks like LangChain and APIs like OpenAI, skills that a good backend engineer can quickly acquire.
Understanding the Roles
Backend developers are the architects of server-side logic, databases, and application infrastructure. They use languages like Python, Java, Ruby, or Node.js to construct APIs, manage data storage, and ensure efficient data processing. Their role is expanding to include implementing AI services.
With the proliferation of 3rd party APIs to serve many needs, today’s backend developers are accustomed to connecting applications to multiple external APIs to solve business problems. Need to process payments — your dev is connecting to Stripe’s API. Need to send SMS messages, they are using Twilio. Now, LLMs are just another 3rd party API with OpenAI leading and many other players chasing them including all of Big Tech. It’s worth noting that the LLM APIs are then hitting LLMs so instead of the reliable behavior of things like a credit card processor, things like hallucinations can (and will) occur.
AI/ML engineers specialize in artificial intelligence and machine learning. They leverage Python, R, or specialized ML frameworks to develop and implement ML models and algorithms, necessitating robust mathematical and statistical acumen. However, their work often requires backend skills for deployment and scaling.
When to Choose a Backend Developer
Opt for a backend developer when building traditional web applications, developing RESTful APIs, optimizing database performance, implementing authentication systems, or scaling existing applications. Increasingly, they’re also the right choice for integrating pre-built AI services.
Example: Leveraging Existing AI Services
Using OpenAI API or similar services often requires implementation rather than AI expertise. A backend developer with API integration experience may suffice, making them the more common need. Creating wrappers or custom interfaces is primarily a backend development task, demanding strong API design skills. Fine-tuning existing models may need limited AI/ML expertise but can often be handled by a backend developer with some AI knowledge.
When to Choose an AI/ML Engineer
AI/ML engineers are indispensable for implementing predictive models, developing natural language processing systems, creating computer vision applications, building recommendation engines, and optimizing complex algorithms for pattern recognition. Choose them when you need to create or significantly customize AI models.
The Gray Area
The blurring lines between these roles create a substantial gray area. Data-driven applications often require both skill sets. Intelligent web services might demand hybrid expertise, while scalability concerns for AI models may necessitate backend proficiency. This overlap is where many companies struggle to define their needs accurately.
Example: Training AI Models
Custom model training is necessary for unique problems or when proprietary data could provide a competitive edge. This process demands significant computational power and large, high-quality datasets. Data preparation and cleaning, often the most time-consuming aspect of AI projects, requires collaboration between data engineers and AI/ML engineers. However, many companies overestimate their need for custom models.
Making the Decision
When deciding between an AI/ML engineer and a backend developer, consider the following factors:
- Core Value: Determine if your project’s value primarily stems from intelligent predictions, pattern recognition, or more traditional data processing and management.
- Primary Concerns: Evaluate whether your main focus is on data management, API development, application architecture, or specialized AI capabilities.
- Project Scope: Assess the need for both traditional backend services and AI capabilities. Often, projects require a mix of both.
- Data Situation: Examine your current data infrastructure, quality, and volume. This can significantly influence whether you need specialized AI expertise or if a backend developer can handle your needs.
- Timeline and Budget: Remember that AI/ML projects typically require more time and resources. A skilled backend developer might meet your needs more efficiently and cost-effectively than a specialized AI/ML engineer.
- Existing Infrastructure: Consider your current tech stack and how new developments will integrate with it.
- Long-term Goals: Think about your future scaling needs and how your choice now will impact later developments.
The blurring lines between these roles add complexity to this decision. Many developers are cultivating hybrid skills to bridge this gap, and backend developers are increasingly equipped to handle AI integration tasks. In today’s dynamic AI landscape, particularly with the advent of Large Language Models (LLMs), the skills needed often align more closely with those of a good backend engineer who can quickly adapt to new technologies and APIs.
When in doubt, don’t hesitate to consult with tech leads or platforms like Gun.io. Making the right choice can significantly impact your project’s timeline, budget, and ultimate success. And remember, in many cases, that choice might be a versatile backend developer with the ability to integrate AI services effectively, rather than a specialized AI/ML engineer.
The post Do you need an AI/ML Engineer or a backend developer? appeared first on Gun.io.