How a Top Freelancer on Upwork Solves Complex Problems With AI and Deep Learning

Richard Alexander has always been passionate about problem solving. He has a background in physics that shapes how he approaches problems and has helped him find success as a freelancer in the fields of artificial intelligence (AI) and deep learning.

“The great thing about physics is you break everything down into its very smallest elements and take a problem one piece at a time,” said Alexander. “And it turns out, this approach is a pretty good way to tackle just about any problem.”

Alexander spent much of his career in scientific research, including roles at Cerner, a health information technology company, and the Stowers Institute for Medical Research. During his time at Cerner, Alexander learned about programming and developed his first business automations. At the Stowers Institute, he focused more on AI and deep learning, specifically computer vision and natural language processing.

After more than a decade at the Stowers Institute, Alexander started freelancing on Upwork and ultimately decided to leave his full-time position to focus on his growing business. He now specializes in AI and deep learning, creating custom neural networks to automate business processes and solve client problems based on their specific business domains.

Transitioning from full-time work to freelancing

While Alexander enjoyed his job and collaborating with colleagues at the research institute, after about 13 years, he reached a point when he wanted to explore ways to earn additional income.

“I looked for opportunities online, found Upwork, and started a freelancing business, which has been fantastic,” said Alexander. “I completed projects on the side for a few years while continuing to work my full-time job. I eventually got one really good client and in 2020, after a year or two of working with them, the earnings essentially replaced my income from the research institute. At the time, making the transition from a full-time job to freelancing was the natural progression.”

An important benefit of freelancing that Alexander has experienced is increased flexibility and work-life balance. He has four children and can now eat lunch with his family during the workday. Alexander also saves time that would have otherwise been spent commuting, and he’s always home for dinner. This increased flexibility also allows him to better enjoy his hobbies, including the occasional long lunch to sneak in a quick tennis match with friends.  

‍

Another advantage of working as an independent professional on Upwork is earnings potential. Alexander has opportunities to identify and reach out to prospective clients, as well as to receive project requests from clients on the platform based on his skills and experience.

When Alexander first started on Upwork, he focused on website tasks and back-end development to begin building a client base, before eventually moving into full-stack web development and deep learning.

“I really focused on building my business and worked on various projects to get more reviews and ratings,” said Alexander. “Now, I'm at the point where clients come to me and they're asking me for work. I don't have to go and search for business and it's ideal.”

Neural network and deep learning project examples

The terms “neural network” and “deep learning” are sometimes used interchangeably, and understanding the difference between them is important.

A neural network is a type of artificial intelligence that helps computers make decisions and complete tasks in a stepped, iterative way that was inspired by the human brain. Deep learning is a subcategory of machine learning that focuses on training neural networks to learn to do a task by example. Rather than coding the exact solution, a dataset is curated that contains both the question and the answer to help the neural network learn the task. Some common use cases for deep learning and neural networks include text, audio, video, and image classification and identification.

Most deep learning projects are developed using Python, a common programming language used to build websites and software, automate tasks, and complete data analysis. Alexander obtained Upwork’s certification for Python development. His projects typically focus on Python programming using either TensorFlow, a deep learning platform developed by Google, or PyTorch, a deep learning framework by Meta, along with other niche tools for image or text processing tasks.

Alexander’s core focus area involves creating custom neural networks for clients, specifically in the areas of computer vision and natural language processing.

Here are two examples of recent projects Alexander has worked on involving deep learning and neural networks:

• Architecture blueprints. Alexander recently completed a project for a group of architects. In architecture, blueprints are often so massive that parsing through all the data can be both challenging and time-consuming. For example, while a house may have two external doors and a few windows, larger buildings may have hundreds of windows. During construction, blueprints are passed around from one person to the next, and small but important details like windows can be miscounted or overlooked. Using neural networks and deep learning, a computer can track these key details, which leads to improved accuracy and efficiency.
• Legal documents. Similar to the architecture project, Alexander is working with a lawyer to streamline a large folder of legal documents. With neural networks and deep learning, PDFs can be organized into a more readable format and categorized into specific folders based on what each document contains. This automatic categorization saves a significant amount of effort that would have otherwise been spent manually searching through the documents.

Based on Alexander’s experience and the projects he’s completed, the top advantage he sees with AI and neural networks is time savings. “The largest and easiest benefit to cite is reduced man-hours,” shared Alexander. “We're going to save a ton of manual effort by having a neural network complete tasks and processes that maybe a person can do, maybe a person can't do, but either way the process is going to be a lot faster and ultimately more affordable than having someone complete the task. Labor is expensive. Computers are cost-effective in comparison.”

Advice for clients interested in AI and deep learning

Because AI is a new and evolving field, potential clients with limited knowledge will approach Alexander wanting to incorporate AI to automate processes. However, in some cases the better solution is traditional programming. Alexander has the expertise to identify the right solutions for clients’ problems—including whether or not AI is the best approach.

“The old adage about accessing information on the internet is that it’s like trying to take a drink out of a fire hose,” shared Alexander. “This is the case with AI because so much information is available and new developments are released each day. There's an overwhelming amount of information, as well as multiple fields and areas of research in AI to aggregate and understand. Because of this, leveraging a professional who is savvy in how AI works is one of the best ways to speed the process and determine the best approach to a specific problem.”

When it comes to neural networks and deep learning specifically, Alexander said the most important consideration when working with clients is having a shared understanding of what data is available for the project.

“The dataset that the AI is going to train on is the entire product and the neural network can't learn anything that's outside of the dataset,” explained Alexander. “It's really simple. If I give you a book to read and then ask you questions that aren’t in the book, you won’t have the answer. Similarly, neural networks learn tasks based on a specific dataset and can only learn from what’s available. The single best step we can take to drive success with deep learning projects is creating a strong dataset from the very beginning.”

Why human intelligence is essential to AI

While the demand for AI has increased significantly and the field will continue to grow, Alexander emphasized the importance of human intelligence powering artificial intelligence.

“Artificial intelligence is in my mind an oxymoron,” said Alexander. “It's a contradiction of terms. No intelligence is artificial. What this means is any intelligence that we provide to a machine is infused with some subset of our intelligence. We as individuals have a massive role to play in AI because we create it. There's no AI without somebody to make AI.”

Because humans are essential to the success of AI, Alexander also believes that job losses related to AI won’t be as impactful as some predict. “If you look back historically at most major innovations, such as railroads or cars for example, there have always been discussions about humans losing jobs. However, in many cases, innovations have created more jobs. People may lose jobs we already know about, but these jobs will be replaced by new jobs that may not even exist yet.”

Staying up to date with the latest AI terms and developments is essential as AI continues to become more widely adopted. As Alexander pointed out, AI research is released at a blistering pace, which can be daunting. To learn about key updates, he keeps track of deep learning conferences that take place throughout the year. Research papers related to each conference are typically released about a month in advance. Alexander reads the latest research to learn about areas of interest and identify new ways to continue driving the best possible outcomes for clients.

Engage a proven expert for deep learning and AI projects

Alexander is a skilled deep learning and AI professional and an Expert-Vetted (top 1%) freelancer on Upwork. He’s passionate about solving clients’ problems using deep learning and has decades of experience in the creative process of taking an idea from conception to production, as well as more than 20 years of coding and web development experience.

If you’re interested in engaging Alexander for a deep learning or AI project, take a look at his Upwork profile, book a consultation, or invite him to apply to one of your jobs.