When you click on a sponsoring school or program advertised on our site, or fill out a form to request information from a sponsoring school, we may earn a commission. View our advertising disclosure for more details.
According to Chris Sullivan, an aerospace engineer formerly at NASA, computer vision—a subfield of artificial intelligence—is in high demand for 2018. Broadly defined, computer vision is the ability of a machine to process, classify, and distinguish between images, digital or real. While this field has its roots in the 1960s when university researchers attempted to develop AI systems mimicking human vision, it’s only recently that companies across multiple industries have called for skilled professionals in this area.
The applications of computer vision are broad, including uses in virtual reality, robotics, navigation (e.g., self-driving cars), modeling (e.g., constructing 3-D reproductions from 2-D images), manufacturing inspection, and the rapid organization of information, among others.
Engadget (Dec. 2016) reported that a neural network system at MIT’s Center for Brains, Minds and Machines (CBMM) not only could recognize faces, but also had autonomously developed a method to recognize rotated portraits. This worked as long as control pictures were rotated at the same angle. This research has broad commercial applications for systems such as “face unlock” for phones and computers, as well as auto-tagging people in photos on Google, Facebook, or Apple. Law enforcement is also eager to use computer vision to sort through countless DMV photos and security camera recordings to pinpoint the locations of criminals.
While the specific requirements vary by application and company, there are some commonalities. Many positions call for candidates with at least a bachelor’s degree in software engineering (or a related field), as well as experience in areas such as machine modeling, tracking, and detection. Experience with C++ and knowledge of concepts such as Bayesian filtering, prototyping, machine learning, and 3D geometry are also typically preferred.
While virtual reality development has been around since the Sensorama (i.e., experience theater) made its appearance in the 1960s, it’s only recently that the prospect of multisensory artificial experience has come into vogue. In fact, Vice (July 2015) reported that “The Feelies” was a groundbreaking VR movie experience which offered not only visual stimulation, but olfactory and tactile experiences as well. VR is an especially hot field for 2018 since it has wide-ranging applications in therapy (e.g., treating PTSD), training for athletes, tourism, medicine, education, real estate, and of course, entertainment. In an October 2016 interview with OEP, Stanford scholar Dr. Bireswar Laha stated, “I think CS, driven by the wave of VR, is becoming increasingly interdisciplinary. It’s now hard, or almost impossible to limit any CS project just to CS—collaborations between CS people and researchers and scientists from medicine, natural sciences and humanities are now very common.”
Again, the responsibilities vary by industry and corporation, but most companies call for at least a bachelor’s degree in computer science, software development, or a related field, as well as experience with prototyping, building simulations, and various programming languages (e.g., Scala, C++, C#, Python, Java, Bash). Preferred qualifications may include experience with versions of Unity, Unreal Engine, and/or Blender.
In July 2015, Wired journalist Andy Greenberg allowed two car-hackers to remotely assume control of his Jeep Cherokee through its wifi-enabled UConnect infotainment system, a situation which blew the lid off concerns about vehicle systems access and safety. Whether self-driving or traditional, suddenly cars were seen as potential hacking targets, forever changing the employment landscape in the field of cybersecurity.
In the aforementioned interview with OEP (Oct. 2016), Dr. André Weimerskirch—the Vice President of Cybersecurity for E-Systems at Lear Corporation—stated that, “Security for cyber physical systems, such as vehicles, is new and not well understood…In vehicles even a single security breach that endangers passengers’ safety is not acceptable. We will need low-cost security solutions for vehicles that are resilient to security compromises.”
For self-driving cars, the emergent field of automotive cybersecurity combines knowledge of not only AI fundamentals (e.g., deep learning, computer vision, machine learning, parallel algorithms, etc.), but also programming techniques fundamental to protecting the integrity of digital data. In addition to at least a bachelor’s degree and facility with programming languages (e.g., C/C++), preferred qualifications for this field include experience working with open source projects, integrating V2X and ADAD functionality, and algorithm development.
Dr. Weimerskirch concluded his OEP interview by saying, “There are several ways to start working in this space, and I think a good one is to take courses in cryptography and applied data security, and also learn a bit about automotive electronics. Then it’s a good idea to get in touch with a specialized group that you can find at a few universities such as the University of Michigan or the University of Tulsa.”
Bitcoin stunned the financial world when it was introduced in 2008: a cryptocurrency with a discrete number of units and all transactions available in a public ledger called a blockchain. One of the benefits of this technology is that each transaction can be traced by anyone with the software, thereby increasing financial transparency and discouraging fraudulent transactions.
The Financial Technologies Forum (Dec. 2016) speculated that blockchain could decimate back-office processing positions (i.e., workers in banking, payments, and securities) since it eliminates intermediaries in financial transactions. Also referred to as “distributed ledger technology” (DLT), blockchain was very recently adopted by the Depository Trust and Clearing Corporation (DTCC), which plays an important role in recording all stock and bond trading across the US. Some people have likened this move to the historically disruptive power of robotics on manufacturing, and time will tell what impacts blockchain has on financial services worldwide.
This is one emergent career where years of software engineering (as opposed to a formal degree) may qualify someone for an entry-level position. Also desired is experience with cryptography, REST-compliant web services, distributed systems architecture, open-source projects, RDBMS or NoSQL databases, hyperledger or ethereum, NodeJS, and programming languages (e.g., GO, C++).
Wired (May 2016) published a piece summarizing the essence of deep learning, a discipline closely related to computer vision and AI. Deep neural networks are fashioned after the neocortex of the brain to analyze vast amounts of data and recognize patterns. This technology has far-reaching implications not only for identifying faces in photos, but also for language translation, adding sound or color to silent or B&W movies, classifying objects, and even generating content, to name a few uses. (On that note, how do you know this isn’t written by a machine?)
A large part of this field is experimentation, the trial and error affiliated with designing the big breakthroughs in AI, ideally by “automating some of the heavy lifting” and unleashing the power of machines to be autodidacts and self-testing systems through complex algorithms. MIT Technology Review (Jan. 2017) recently reported on a promising machine-learning technique called generative adversarial networking (GAN), which has two systems: a data-creator and a data-regulator (i.e., one that distinguishes between real and fake data). GAN is expected to have applications in generating video-game scenery and clarifying overly pixelated video footage, among others.
Compared to the previous four positions, deep learning engineers and scientists typically require more formal education, ideally a master of science (MS) or a PhD, due to the field’s solid grounding in scholarly research and theory. Additionally, many specialized deep learning positions call for candidates with experience developing algorithms for various systems (e.g., image and video analysis), prototyping, feasibility studies, open source projects, machine learning, object detection, statistical modeling, pattern recognition, applied math, and programming languages (e.g., C++).
These are just five of the exciting fields ripe for exploration, areas which existed mainly in university laboratories in 2012. Journalist Scott Dadich leaves us with a pressing question: “We’ve heard from folks like Elon Musk and Nick Bostrom who are concerned about AI’s potential to outpace our ability to understand it. As we move forward, how do we think about those concerns as we try to protect not only ourselves but humanity at scale?”
After graduating from UC Berkeley, Jocelyn traveled the world for five years as an English teacher and freelance writer. After stints in England, Japan, and Brazil, she settled in San Francisco and worked as a managing editor for a tech company. She’s recently become interested in the underrepresented groups in STEM fields and published a piece on Women in Engineering. When Jocelyn isn’t writing about college programs and professors, she satirizes global politics and other absurdities at Blore’s Razor (Instagram: @bloresrazor).