Many Routes to an AI Career

Artificial intelligence technology grabs headlines every day, with its developers touting ever more sophisticated capabilities such as generating videos, duplicating voices, and translating languages in near real time.

Behind the scenes, AI has quietly pervaded such everyday activities as screening resumes, analyzing apartment rental applications, and reviewing medical procedure results. AI also plays an increasing role in our daily lives as it combines with sensors, robotics, and geolocation technologies to provide health monitoring, GPS guidance, and intuitive personal assistants. Investors are boosting the stock market fortunes of AI innovators and companies that provide cloud computing and AI infrastructure, despite concerns that the technology can “hallucinate” — misinterpret images, repeat discriminatory stereotypes, and fail to spot harmful images — among other challenges. (Read “The Challenges and Opportunities of AI Part 1: Bias in AI.”)

AI’s reach and the controversies surrounding it are nothing new for three women who studied or worked in engineering and have won acclaim from Women in AI (WAI), a global network of women AI professionals working toward inclusive AI that benefits a global society.

These AI leaders said, in separate interviews, that in their work with AI they use the very same skills that are critical to engineering — mathematics, computation, systems design, safety analysis, and efficiency — to root out threats, inaccuracies, and unfairness that would undermine the highest and best standards of AI systems. They see AI, under the proper guidance, as a way to free people to explore their creativity and interests, and they view a career in AI as a challenging vocation that needs more women leaders to help steer its future.

AI innovator

Haydé Martínez, head of data and AI at the software company Wizeline, initially studied to be an industrial engineer. As a Mexican ambassador for WAI, she founded De Cero a Ciencia de Datos (Zero to Data Science), a programming course focused on data science and machine learning.

Recognized by WAI as an AI Innovator of the Year in 2023, Martínez will graduate in May with a master’s in data science from ITESO, the Instituto Tecnológico y de Estudios Superiores de Occidente, a Jesuit university in Guadalajara, Mexico. She proves that taking an unorthodox career route needn’t thwart one’s ambitions.

Martínez grew up and still lives in Guadalajara and remembers as a child sitting next to her dad, a civil engineer, as he worked at a home computer. He subscribed to computer magazines that came with CD-ROMs of programs like Lotus Notes and game demos, which she enjoyed. She received a toy computer of her own when she was just 5 years old, and said, “I would carry it everywhere with me as my best toy. It made me feel like my dad.”

After her mother transferred her to a new elementary school, Martínez was bullied and ostracized because she was “very smart, always aiming to have the top grades, and I was different,” she said. “I just wanted to read, study, and draw.”

Despite this, she never lost her drive. She started working as a door-to-door survey-taker at age 15 and took high school classes at night so she could work full time as a receptionist at 17. At 20, Martínez began contract work at Hewlett-Packard as a business analyst, doing data capturing and cleansing and other repetitive tasks that are now automated.

“The tools and knowledge are there for anyone to access. This is your chance to create, to influence how AI operates, and to write the rules of its behavior.”

– Bhaktipriya Radharapu

After taking a full-time internship at Intel, she studied for a bachelor’s degree in computer systems from 7 to 11 p.m. each weeknight after working 9 a.m. to 6 p.m. At Intel, she was one of only 40 women out of 600 employees. She said colleagues from her team let her know several times, in different ways, that she didn’t belong — at Intel or in the tech world in general. “I cried so much,” she said. “I came home, and my brother said, ‘They want you to cry. They want you to not succeed. ‘You have to build a shell around your heart and in yourself and throw away those comments.’”

That kind of family support provided a big boost, and in Martínez’s second year at Intel, she found a woman role model and a project in software validation that together gave her the confidence she needed.

She graduated college at age 24, and then won a scholarship to study machine learning and AI at the Kanazawa Institute of Technology in Japan. After completing the program, she returned to Mexico and landed a machine learning lead role as the second employee of the startup tech firm Rever.

“I did hiring, talking to investors, cleaning the bathroom — I did everything,” she said. She was asked to speak before 100 Mexican investors when she was 27 years old. By the time she left Rever, Martínez had given 30 talks at a variety of events and realized she was “good at talking.”

During this time, she also started De Cero a Ciencia de Datos to help expand the technological skills of people in the Guadalajara area and throughout Mexico. In 2018, she partnered with the Jalisco Institute of Information Technologies to reach a larger audience, and her company now counts more than 140 graduates. “I love it,” Martínez said. “It’s completely mine.”

She noted that working in AI involves math, programming, and other complex subjects, but that doesn’t make it exclusionary. “We humans are able to learn anything we want,” she said. “Our brains are made to learn five languages, five sports, and on and on. I want more people to belong to the tech world. We need more equity and inclusion in tech.”

Martínez had to overcome two unexpected setbacks: a 2018 car accident that left her in a coma for four days and required her to have reconstructive surgery and her father’s sudden death on May 9, 2021. She persevered, and in July 2021, began a job as a senior data engineer at Wizeline, a technology services provider that builds digital products and platforms to speed its clients’ time to market. She was promoted to data technology program lead and, in May 2023, was named director of data and AI. She now leads a team of more than 100 engineers, scientists, and data analysts.

AI role model

Bhaktipriya Radharapu, a software engineer who works as the technical lead for responsible AI at Google in Mountain View, California, builds machine learning infrastructure to make AI models fairer and more robust. She was named the WAI 2023 Young Role Model of the Year and led Google’s inaugural Summer of Code for Responsible AI program for aspiring machine-learning developers.

Radharapu said she found her calling in high school in Kuwait, where she realized that her interests lay in finding, building, and creating logical outcomes, especially using such programming languages as C++. She credits her open-source contributions to the Linux kernel — a computer program central to an operating system — with helping her gain industry recognition, and as a key driver in her early career with Google as a paid intern.

“I tell young people why they should get involved in and contribute to open-source,” she said. “With my first 50 to 100 patches, I was just trying to learn how to work with the code base. It was free. It was skill learning. And mentors online gave me help, too.”

Radharapu credits her grandfather, an electrical engineer, with serving as a mentor and helping her improve her math skills while she was growing up. He used gamification and real-life questions to make learning math fun.

“He brought in the neighbors and my cousins to figure out questions like, ‘What’s the most direct route to the ice cream shop?’” Radharapu said. “It was not a situation in which I had to get it right. He rewarded me when I made progress. His goal was to go through my entire math curriculum while I was on vacation.”

“AI, to me, is there to augment skills we already have. We can get to decisions faster.”

– Bhaktipriya Radharapu

Radharapu attended college in India at the Indian Institute of Technology Hyderabad, majoring in computer science. While there, she was invited to Paris to research the use of AI to help doctors diagnose cancer. She worked on scalable and efficient algorithms and systems for the segmentation and classification of pathological patterns in 2D and 3D medical imaging data.

Radharapu was then invited to conduct research with the Massachusetts Institute of Technology Computer Science and Artificial Intelligence Laboratory.

From there she began a career as a software engineer, focusing on the deployment of deep learning models on Android devices. The role involved the challenge of engineering Google’s best models into tiny, efficient versions that could fit on phones while ensuring privacy by running locally — without sharing user data. This concept is a key pillar of responsible AI, she explained, and it inspired her to transition into the field of ethical AI development. This move allowed her to explore other aspects of responsible AI, such as safety and fairness.

“AI, to me, is there to augment skills we already have,” she said. “We can get to decisions faster.”

Radharapu’s responsible AI team within Google recently stress-tested the company’s Gemini AI model against a variety of attacks and worked to improve safety and fairness practices in Google’s Workspace, Ads, Maps, Cloud, and Search applications.

AI volunteer

Ivete Sánchez Bravo — a project manager at Centro de Investigación en Matemáticas, a Guanajuato, Mexico-based scientific research institution — won recognition in Forbes México in October 2023 as one of the 20 AI leaders in that country. She also earned praise as a WAI 2023 AI Volunteer of the Year.

Sánchez Bravo, who grew up in Mexico, studied computer systems and earned a master’s degree in computer science from CIMAT. She came to AI through her love of math, which she discovered in secondary school.

She is originally from Ecatepec, in the state of Mexico, an area with one of the highest levels of inequality in the country. Sánchez Bravo’s parents, originally from Puebla, migrated from their hometown to Ecatepec in search of better opportunities. She belongs to the first generation in her family to pursue a bachelor’s degree, which she earned from Escuela Superior de Cómputo I.P.N. in Mexico City.

Driven by her lifelong dedication to mathematics and problem-solving, she chose to pursue engineering as a career despite the scarcity of role models and the fact that she was one of few women in the field where she lived. She now aims to inspire young women to be part of the AI community and has served as a city lead for WAI, representing Bajío, in Mexico’s west central agricultural region. In this role she works as a volunteer to close the gender gap and raise awareness of the need for gender diversity in AI. She also serves as a WAI community operations manager and is a founder of the nonprofit Mexican Women in Data Science.

Sánchez Bravo also served from 2019 to 2022 as the manager of an AI consortium for eight public research centers affiliated with CONAHCyT, Mexico’s National Council of Humanities, Sciences, and Technologies.

A positive outlook

All three of these AI trailbreakers view AI as a tool for good — but only if properly used.

Martínez is an optimist who sees AI as a tool that will complete repetitive tasks so that people can evolve to more intellectual pursuits. “It will happen in the same way we adapted to the internet and to smartphones,” she said. “We will normalize it.”

Sánchez Bravo said she believes it is necessary to regulate AI applications, especially those in Mexico. For example, she said, the algorithms that prescreen resumes for certain positions and generate health risk indicators for patients based on historical information should be considered for oversight, she said. Some companies have internal policies related to ethical risk assessments and transparency in decision-making, but not all have these elements, she said.

Some AI experts say companies should be required to do their own “bias audits” and make the results public, but industry leaders argue that this would expose their trade secrets. Meanwhile, governments worldwide are behind in legislating AI, which creates special concerns in countries that use state surveillance systems and other technologies that can violate people’s privacy when they are misused. All of this still needs to be addressed, Sánchez Bravo said.

Radharapu agreed there is a great deal of work to be done “trying to find an alignment between [AI] being helpful and being harmless. Every big tech company has its own safety and responsibility department, [but] they’re all less than 3 years old,” she pointed out, and require refinement and maintenance.

Now is the best time for young and aspiring engineers to seek out opportunities in AI, Radharapu said. “This is a very exciting time to be in the field. The tools and knowledge are there for anyone to access. This is your chance to create, to influence how AI operates, and to write the rules of its behavior. The only thing stopping you would be yourself.”