a human brain blended with a machine brain

Beyond the Hype: Artificial Intelligence & Machine Learning From Pilot to Production

As the Founder and Chief Product Strategist at TechNoch Solutions, Swathi Young helps CIOs solve complex problems by combining business strategies and solutions for Artificial Intelligence (AI). With a background of over 20 years in the tech industry, her most recent work is centered around finding organizational strategies for increasing needs in the AI industry.

Save Your Seat at Cyber Bytes

“I’m passionate about finding solutions for complex problems,” Young said. “Especially with regard to ethical design of AI applications as well as data privacy.”

Young spearheads a weekly newsletter curating information on AI and machine learning and leads a community of over 2,000 members in the DC area focused on emerging technologies, organizing events on AI and blockchain.

Getting ahead of the hackers with AI and Machine Learning.

“As with innovation, even more so with tech, everything comes with good and bad,” Young said. “It is also true of AI and machine learning.”

Young explained that when it comes to cyber challenges, it’s not only important to think about how AI and machine learning can be used to detect hackers, it can also be used by hackers to remain undetected. As Young puts it, “It’s a double-edged sword.” She challenges cyber warriors to continually ask the question, “Am I up-to-speed in getting ahead of the hackers?”


In keeping pace and getting ahead, Young acknowledges the realities that cyber professionals and corporations face. “Technology infrastructure is often a downfall,” Young said. “It’s unlike traditional software engineering infrastructure,” and, for most of the companies Young encounters, it is in need of updating.

Understanding the problem to find the AI solution.

The first step Young consistently suggests is to start with the business goals in mind. The key question she asks is: “What are some of the cyber problems you are trying to address?” Young also suggests being very specific about the reasons why. The next step is training.

“This [training] is very important in meeting the future demands of the workforce,” Young said.  

Young outlined top three limitations of AI and Machine Learning:

Ethical implications.

“Too often engineers don’t stop and think about the use of AI,” Young said. “It’s unlike traditional IT and software engineering because of the significant potential to misuse it.”

She explains that there will always be ethical implications to consider. Even autonomous vehicles need to answer the question – avoid a pedestrian and sacrifice the driver of the car or hit the pedestrian to save the driver? Another example she gives is with regard to the use of space drones as weapons. These questions require conversations with individuals who are willing to be really thoughtful and engaged. It’s a conversation that often leads outsourcing for expert advice in the realm of neuroscience and psychology, as an example.  

Biased data.

Since AI and machine learning is fundamentally built on top of data, and we have 188 cognitive biases as humans, those biases will sometimes creep in,” Young said.

She shared an example regarding facial recognition that was faulty in its design. The system failed to recognize whether or not a Chinese person had opened or closed eyes. Young also cited the infamous example in the Fall of 2018 in which Amazon had to bring down its HR recruiting tool because it was proven to be biased against women. The historical data in the tool did not include women.

Data privacy.

“We are now familiar with many of the issues regarding data privacy thanks to Facebook and Cambridge,” Young said. “There has also been interference in our election process. Not all folks are aware about rights, intrusion and use without knowledge. It is actually used to manipulate indirectly what you are thinking.”

Young used Facebook ads as an example. There is a whole philosophy about “nudging” and how it impacts thoughts and actions. She said that very recently, Google mentioned the need for focusing on data privacy and enabling common people to adjust their settings. “The steps to do so, however, are so complex it took me 15 minutes – as a technologist – to find the settings.” Young compares the race to AI to the nuclear arms race. But instead of physical attacks, there is a use of AI and data to manipulate thoughts—a trend that will only increase over time with the collection of more and more data.

Action steps to find a successful AI solution:

If attendees have one takeaway, Young hopes it would be that AI has come a long way in the last couple years and has gone beyond the hype and hysteria they have read about.

“Don’t be afraid to start tinkering around to uncover the problems,” Young said. “There are tools out there to easily build AI solutions.” The mantra she repeats is to “always remember to begin with the business problem you are trying to solve.”

Here are some suggested action steps to find successful AI solutions:

  1. Understand what your business goals are. What are you trying to use AI for?
  2. Understand that any AI solution depends on data and the 5 c’s of data – how clean that the data is, for one.
  3. Decide whether or not you want to train your internal team to create solutions or partner with a company like TechNoch solutions.

Join us at our next Cyber Bytes Networking Event at 1000 Corporate Drive, Suite 119, Stafford, VA 22554. Save your seat.

The Cyber Bytes Foundation is committed To establish and sustain a unique Cyber Ecosystem to accelerate the development of a strong Cyber workforce and support community outreach programs creating synergy through Education, Research, and Innovation.