ACM-W Rising Star Award Recipient: Dr Manya Ghobadi

ACM-W would like to announce Dr. Manya Ghobadi as this year’s recipient of the ACM-W Rising Star Award! The ACM-W Rising Star Award recognizes a woman whose early-career research has had a significant impact on the computing discipline.

Dr Manya is currently an Associate Professor at the Massachusetts Institute of Technology (MIT) in the Electrical Engineering and Computer Science Department.

Congratulations on winning the ACM Women Rising Star Award. Can you tell us about your journey in the field of computer science and technology? What inspired you to pursue this field, and what challenges did you face along the way?

Thank you for the warm congratulations on receiving the ACM Women Rising Star Award. I’m deeply honored and humbled by this recognition. My journey in computer science and technology began during my undergraduate studies at Sharif University of Technology in Tehran, Iran, where I pursued a Bachelor’s degree in Computer Engineering. It was during this time that I developed a strong passion for technology, driven by the potential of using computation to make a positive impact on society.

My inspiration to delve further into the field came during my graduate studies at the University of Victoria and later at the University of Toronto, where I had the opportunity to research and explore my interests in congestion control and data center networks. These fields, coupled with my fascination for hardware-software co-design, optical networks, and network optimization, have shaped my career and research trajectory.

As I moved through my academic and professional career, I faced several challenges. Adapting to new work environments and cultures, such as during my time at Google and Microsoft Research, presented its own unique challenges. These experiences, however, helped me grow and learn the importance of perseverance, resilience, and embracing diversity.

In my current role at MIT, I continue to focus on my research interests, including networks for machine learning. I hope to contribute to the development of cutting-edge technologies that will revolutionize how we compute, communicate, and solve complex problems.

You initially spent some time in the industry at Microsoft Research and Google before taking up a position at MIT. What prompted you to make the switch from industry to academia? How easy/difficult is it to make the switch?

At both Microsoft Research and Google, I had the privilege of working with some of the brightest minds in the industry. My time there provided me with invaluable insights and hands-on experience, and I learned a lot about cutting-edge technologies and research methodologies. However, as fulfilling as my work in the industry was, I realized that I deeply missed the joy of working with students and contributing to their growth and development.

I always had a passion for mentoring and sharing knowledge, and academia provided the perfect opportunity to combine that passion with my research interests. The decision to transition from industry to academia was driven by my desire to contribute to the development of the next generation of computer scientists and engineers, fostering a collaborative environment where students could learn and grow.

In terms of the ease of making the switch, I would say it depends on the individual and their specific goals and circumstances. For me, the transition was relatively smooth, as I was able to leverage my industry experience and research background to devise next-generation solutions. Moreover, the skills and knowledge I gained in the industry were highly applicable to my new role in academia, and I found that I could effectively use my experiences to guide my research and teaching endeavors.

According to you, what are the most pressing problems in Computer Networks today? How do you think your research is making a difference in addressing these problems, and what impact do you hope it will have?

The most pressing problems in computer networks today are the increasing demand for efficient network infrastructures that can support the growing complexity and scale of new applications, such as machine learning, video calls, telepresence, augmented reality, and healthcare. As these applications evolve, there is an urgent need for scalable, cost-effective networks that provide high bandwidth, low end-to-end latency, and high availability, while minimizing energy consumption.

My research is focused on addressing these challenges by developing next-generation systems for emerging applications. A key aspect of my work involves enabling physical-layer reconfigurability in modern networks to achieve high throughput, low latency, and fast recovery from failures. Utilizing advanced hardware, such as optical devices, I develop network architectures, algorithms, and protocols to optimize resource use, energy consumption, and high availability.

For instance, the ever-growing demand for machine learning-based services has led to a steady increase in the dataset and model sizes of deep neural networks. Although specialized hardware accelerators and software stacks have provided significant speed-up in compute capabilities, large-scale deep neural network models still require enormous computational resources and consume substantial amounts of energy. The overarching goal of my work is to build fast, efficient, and sustainable machine learning-centric systems.

By addressing the challenges posed by the growing complexity and scale of contemporary applications, I hope that my research will provide a foundation for the real-world deployment of dynamically reconfigurable networks. This will enable better support for machine learning applications, improve the performance of network-based services, and contribute to a more sustainable, energy-efficient future.

In one of your most cited papers, “OpenTCP: Rethinking end-to-end congestion control in software-defined networks,” which was written in 2012, you and your co-authors argued for the need for a congestion control adaptation mechanism in datacenter networks. How have the insights you presented in this paper changed or held up since then?

The insights presented in the OpenTCP paper have largely held up over the years, and the need for a congestion control adaptation mechanism in datacenter networks remains as relevant today as it was then. OpenTCP has helped network operators define rules for tuning TCP as a function of network and traffic conditions, leveraging the global network view available at the network controller for faster and more accurate congestion control decisions. This shift in perspective has sparked further research into developing more sophisticated and adaptable congestion control algorithms that cater to the unique requirements of modern network applications and datacenter environments.

Despite these advancements, congestion control is still not a completely solved problem. As network infrastructure continues to evolve and new applications emerge, we must remain proactive in developing innovative solutions to address the challenges of congestion control in increasingly complex environments. This includes taking into account the unique demands of emerging applications that place different strains on network resources.

For example, in a recent paper published in 2022, my group demonstrated that fair-sharing, which has long been considered the holy grail of congestion control algorithms, is not necessarily a desirable property for machine learning training clusters. We showed that for specific combinations of jobs, introducing unfairness can improve the training time for all competing jobs. This finding underscores the importance of re-evaluating traditional congestion control paradigms and designing algorithms that cater to the specific requirements of modern network applications, such as large machine learning training jobs.

Based on your experiences studying and working at esteemed academic institutions like UofT and MIT, what advice would you give to aspiring researchers looking to maximize their potential and make the most of their time in academia?

Based on my experiences, I would offer three key pieces of advice to aspiring researchers looking to maximize their potential and make the most of their time in academia:

1. Cultivate curiosity and passion: Pursuing a research career requires genuine curiosity and passion for your chosen field. Stay open to new ideas and dive deep into subjects that truly interest you. Your enthusiasm will not only fuel your work but also inspire those around you.
2. Embrace challenges and learn from failure: Research is often a process of trial and error. Don’t be disheartened by setbacks; instead, use them as opportunities to learn and grow. Embrace challenges and remain persistent in the pursuit of your goals.
3. Collaborate and network: Engage with your peers and researchers from other disciplines. Collaboration often leads to novel insights and innovative solutions. Attend conferences, workshops, and seminars to expand your network, as these connections can lead to fruitful research partnerships and future career opportunities.

As a role model for young women in technology, how do you think we can encourage more women to pursue careers in this field?

One way to encourage more women to pursue careers in technology is by providing early exposure and education in the field. Introducing girls to technology and computer science at a young age can spark their curiosity and demonstrate the potential for creativity and innovation within the field. As they grow older, connecting them with strong female role models and mentors can offer valuable guidance, support, and inspiration, showing them the diverse range of successful women who have built careers in technology.

Creating inclusive and supportive environments is also essential. By fostering an inclusive culture within academic institutions and organizations, we can ensure that everyone has access to equal opportunities, regardless of gender. This can be achieved through policies and practices that encourage participation in networking and professional development events, which can help develop crucial connections and boost their confidence.

As I reflect on my own journey, I’m grateful for the opportunities and support I’ve received, and I’m excited to continue contributing to the advancement of technology. My hope is that, by sharing my experiences and insights, I can inspire the next generation of women in tech to break barriers and make their own lasting impact on the world.

---