Can you describe your AI journey? 

I started as a Product Manager for a complex financial reporting product and realized how AI technologies such as NLP and computer vision, can transform the financial services sector. This was something I hoped to learn someday. I started specializing in data engineering on the cloud, helping build large data processing pipelines that scale-out on the cloud to churn and process vast amounts of disparate datasets in enterprises. The projects soon evolved into building and deploying AI & ML models at scale which involved the challenging task of ensuring that the pipelines return feedback from models to the system for retraining and further refining the model for improved accuracy. My interest and experience today lie in deploying AI/ML models at scale to service tens of thousands of requests, successfully helping customers deploy AI/ML models to enable efficient services to their customers and employees.

What are the major challenges you faced as a woman in reaching where you are right now?

Having women engineers in early-stage careers is quite common, but as I grew in my career, I realized how difficult it gets to climb the ladder. I had to try harder than some of my male colleagues to build credibility in technical discussions. There were several times I wished I could have contributed but didn’t because it was too much effort to make myself heard. I am thankful for my allies across all genders who have helped me build confidence in myself and overcome the perception challenges I face. They stand up for me ensuring meetings are inclusive and all voices are heard. 

What made you interested in AI?

The ability to derive insights from data and using it to predict outcomes that can help save lives, provide efficient services to citizens, enable differently-abled people to connect well with the rest of us are some of the aspects that drove my interest in AI. AI harnesses the true power of data and this got me super interested in building AI-enabled systems.

What's your area of expertise in AI and why choose that one?

I am not a Data Scientist and my strength was always in building data processing flows and pipelines. In the area of AI, my experience lies in taking a model built by scientists and scaling it to production-grade. For example, ensuring that the model is trained over high quality, accurate and complete data, that the model is able to service a large set of users in the given timeframe, and most importantly, pipelines that feed the model actions back into re-training of the model, improving its accuracy, identifying and reducing any biases etc. 

 Off late, I am also very active in the community driving the ethical use of AI, understanding global regulations, what governments and AI stakeholders are doing to drive responsible use of AI that is not only compliant with data privacy laws but also review the applicability of AI with respect to AI principles of Fairness, Inclusiveness, Accountability, Transparency, Reliability & Security and Privacy.

What's the one thing that you see AI transforming completely? 

I see our ability to communicate with each other transforming with the use of AI, computer vision, voice recognition, natural language translators, voice-enabled chatbots etc. It brings communities of people together bridging the gaps that were there earlier. It has the power to make our world more inclusive – people with hearing difficulties can follow a presentation through automated transcripts, people with low vision can now make use of Smart Glasses to navigate independently. 

Personally too, when my husband hurt his hand and couldn’t type, the dictation feature embedded in Microsoft Office helped him continue working! My mother uses voice-based search to look for news, new recipes, learn new things when earlier she was limited as she wasn’t comfortable typing. In short, it has opened up a completely new world for her. 

Your biggest AI nightmare?

My AI nightmare is that we won’t realize its negative effects unless it has been consistently used over a period of time, by which time the damage can be far-reaching. AI-enabled systems often become black boxes, where one cannot really explain why an AI model made a particular decision. There is a high reliance on these models to make critical decisions right from measuring an individual’s performance to allocating benefits. These models are being trained on historical data which may be biased and could result in these systems causing harm, physical or psychological injuries, denial of service and in some cases impacting human rights. 

 To cite a personal example, both my husband and I applied for a premium credit card with a bank. His got accepted and mine rejected. The issue was that nobody at the bank could explain the reason. They all said a system decides this based on hundreds of/multiple parameters. My husband and I work for the same company, draw similar salaries, are of the same age, live at the same address, so it was difficult for me to understand why mine was rejected. It is totally possible that I may have had a bad credit history but having the ability to explain this in AI models will go a long way in avoiding chaos and confusion.

What's your advice for other women who wants to pursue a similar journey?

My advice is very simple – set aside sometime every day to think, plan and work on learning new technologies and skills that will keep you relevant in this fast-changing domain. Additionally, build a network of diverse allies that will help you skate through the challenges we often face. They not only help you grow at work but also become your friends.