![Rapid engineering: an integrated dream](https://technicalterrence.com/wp-content/uploads/2024/02/Rapid-engineering-an-integrated-dream.jpg)
Image created by me with Microsoft Image Creator
Since OpenAI released ChatGPT to the public, a flurry of online discussions have emerged about their new dream job: Prompt Engineering. It is promoted as “ai-s-hottest-job-prompt-engineer” rel=”noopener” target=”_blank”>The hottest job in aipromising six-figure salaries technology/2023/02/25/prompt-engineers-techs-next-big-job/” rel=”noopener” target=”_blank”>no programming experience required. Enthusiasts describe it as a work of the futurewhere ai” rel=”noopener” target=”_blank”>anyone can win until ai-prompt-engineer-job/” rel=”noopener” target=”_blank”>$335K convincing a cool know-it-all robot to give the right answers. No surprise, Instagram wise men who make moneyYouTube career preachers, and the self-proclaimed oracles of TikTok have been very explicit about it. While this sounds like your dream job, is it really achievable? Let's delve into the job market reality behind the hype to find out.
Analysis of job posting data provides valuable insights into job demand trends, responsibilities, qualifications, and salary expectations. Therefore, I decided to take a look at the data from the ads of the so-called “ai-s-hottest-job-prompt-engineer” rel=”noopener” target=”_blank”>The hottest job in ai”Without speculation or presumptions. I collected 73 data from unique job advertisements recently posted on popular online job posting platforms. Read about my data collection methodology and access the data set. here. While 73 may not be an ideal sample size, it is an integral starting point for our analysis. The initial revelation is sobering: there is a shortage of employers looking for “fast engineers.”
Now, let's look at the data. The most frequently mentioned job title is “rapid engineer.” However, other titles such as “IT Innovation Analyst,” “Freelance ML/ai Engineer,” “Data Scientist,” and “ai Engineer” are also emerging. I created word clouds for the qualifications and responsibilities mentioned in the job descriptions. I don't think word clouds are meant to reveal extraordinary ideas, but they can represent a compact version of the text's important highlights. As you can see, in job ads employers talk about experience in computer science, model development, Python, rapid design, machine learning, large language models, natural language processing, and artificial intelligence more than other things.
1. This is a significantly larger sample size compared to many of those early anecdotal articles that built their entire argument about a six-figure salary without coding from a single job ad.
Next, I used ChatGPT and Claude to summarize the text corpus of the collected advertisements to identify the most important engineering qualifications and qualifications. I did several rounds of prompting with different approaches and then manually checked the data to make sure I got stable and valid results.
Essential qualifications required for the Prompt Engineer job:
- Proficiency in Python programming (2-5 years of experience) including experience with ai/machine learning frameworks such as TensorFlow, PyTorch, Keras.
- Working knowledge of NLP and LLM (2-5 years of experience) such as BERT, GPT-3/4, T5, etc. Knowledge of how these models work and how to adjust them.
- Strong analytical and problem-solving skills.. The ability to think critically, design effective prompts, analyze model performance, and troubleshoot is vital.
- Experience in rapid engineering principles and techniques. such as chain of thought, learning in context, thinking tree, etc. This allows the models to be guided towards the desired results.
- Excellent communication skills, both verbal and written. This is necessary to collaborate across teams, explain technical concepts, and document work.
And the essential responsibilities of specific engineering work are:
- Fast design and optimization: Design, develop, test and refine ai-generated text messages to maximize effectiveness for various applications. This includes using techniques such as transfer learning and leveraging linguistic experience to develop diverse, high-quality prompts.
- Integration and implementation: Ensure seamless integration of optimized indications into the overall product or system. Collaborate with engineers to implement guidance and models in production environments.
- Performance evaluation and improvement: Rigorously evaluate rapid performance using metrics and user feedback. Perform continuous testing and analysis to identify areas for optimization and early iteration.
- Collaboration and requirements gathering: Work closely with cross-functional teams such as data scientists, content creators, and product managers to understand requirements and ensure guidance aligns with business objectives and user needs.
- Knowledge sharing: Document rapid engineering processes and results. Educate teams on immediate best practices. Stay up to date on the latest ai advancements to deliver innovation
It is fair to say that the “no programming experience” premise of the so-called “hottest ai job” is far from reality, as the most in-demand skills in the rapid engineering market are mastery of programming and experience in NLP and LLM. And they're not talking about Micky Mouse's programming skills, they're looking for experts who are familiar with machine learning and artificial intelligence frameworks. Employers not only demand “familiarity” with LLMs and coding, but on average, they look for experts with 2-5 years of experience working with structured and unstructured data, coding, NLP, ML, and ai.
Reading the main responsibilities makes it clearer why this job requires such a high level of programming and LLM skills. Rapid engineering, as a professional job, is not about sitting behind a computer and playing with generative ai models to give it the right answer. It is about building business information systems that optimize inputs, integrate them perfectly with other information systems and products, and deliver values to users and customers. In other words, companies are not looking for someone who can chat with ChatGPT, they want to hire experts who can optimize GPT-like models and integrate them with their own products.
Analysis of job advertisement data on degree requirements indicates a preference for technical training in computer science, mathematics, analysis, engineering, physics, or linguistics. A bachelor's degree in computer science or a related field is usually required, and more advanced degrees are preferred or required for high-level positions. Salaries are very different depending on responsibilities and seniority. It can be as low as 30,000 and as high as half a million dollars per year. On average, job ads with salary information pay between $90,000 and $195,000 per year.
Despite initial enthusiasm, questions have been raised about the viability of rapid engineering as a dream job. As Wharton School professor Ethan Mollick wrote in a twitter post Last year, “the fast engineer is not a job of the future” because “ai becomes easier” and smarter at interpreting basic indications. A month ago Coursera published a well thought out professional guide for rapid engineering (see also this). It seems that the initial ai generation fad is slowly fading and we are in a better position to understand the current state and future trends of ai. Do not misunderstand. The quality of the results of Generation ai depends largely on the inputs. Learning to use and interact with these complex models is becoming an important skill for almost everyone. There is a growing number of scientific studies suggesting that a systematic approach to indications can significantly improve the outcome of these models (see 1, 2, 3, 4, 5, 6, 7). However, “rapid engineering” is not (and never was) the dream job some people wanted it to be. Without significant experience in programming, natural language processing, machine learning, product development, and software integration, no one will pay you a six-figure salary simply for turning ChatGPT into a correct answer.
The present and future of warning engineering and Generation ai applications appear to be influenced by two important trends: First, as Ethan Mollick mentioned, Generation ai models are becoming more adept at to generate good results from simple and unsophisticated ads, perhaps in a similar way to how the Internet works. Search engines have gotten better at returning more relevant results from simple search queries. Second, Gen ai models are becoming increasingly integrated into enterprise products, services, and platforms. This adaptation is crucial to the success of the ai economy. Therefore, knowing how to optimize, tune, customize, and integrate Gen ai models with current information systems and products is and will continue to be a valuable skill set. That's why in today's job ads, there is a high demand for programmers, system designers, and those who can collaborate with other members of the product development team.
Mahdi Ahmadi He is a clinical assistant professor in the Department of Information technology and Decision Sciences at the University of North Texas, where he teaches data mining, business intelligence, and data analytics. My main area of research is the application of machine learning and data mining techniques in companies. I also provide consultation to businesses, higher education institutions, and nonprofit organizations on their data analytics issues.