As an electrical and electronics engineering graduate, Vaidheeswaran Archana enjoys experimenting with a variety of machine learning and electronics projects. She recently joined the Intel® Software Innovator program and currently works as a deep learning researcher at Saama Technologies in Chennai, India. Archana also has a passion for knowledge sharing and empowering women to pursue work in STEM fields.
What got you started in technology?
The seed of discovery and constant innovation drives every technologist or researcher. For some, there is a single moment that changed their perspective of the world. But for me, appreciation for technology took root in the form of experiencing its change in our day-to-day lives. As a kid born in the ‘90s, I have seen the way a technology as simple as a mobile phone evolved in the past 25 years. My curiosity was piqued when I saw how technology transformed people’s lives. It made me realize its endless opportunities to solve everyday problems.
In my second year of undergraduate study, my curiosity expanded beyond my academic education. My fascination for empirical applications of engineering caused the need for a relentless search for interdisciplinary research projects. During this period, I met some of my peers who were also interested in expanding their ideas into creative projects. We wanted to start a place involved in cutting-edge research and innovation. Thus, we began the first student-run research lab in India: Next Tech Lab, SRM Institute of Science and Technology.
Slowly, we as a team started to implement our own ideas, failed in executing some bizarre ones, and learned to improve on them. Eventually, we expanded our own knowledge base by deriving ideas from students in different departments. This helped us understand and solve real-life problems. Within two years we grew from a team of 10 members to 200. Personally, I was able to mentor over 30 students in the field of electrical and electronics engineering.
Tell us about your experience as a woman working in technology.
My parents have always been very supportive of my ideologies and ambitions. Their encouragement has shaped me to be an independent researcher. They always taught me to value my knowledge and skills beyond anything. I studied in an all-girls school and never looked at my gender as a liability.
I took up electrical engineering in college. When I walked into my first class, I noticed there were only seven girls and 67 boys. I was flabbergasted by this! I couldn’t understand this irrational divide. After all, most of the girls in my class in school took up engineering. It was far worse when I noticed our campus job placements and saw fewer girls taking jobs as engineers. I vowed to not be the same and wished to do something about it. Next Tech Lab helped me raise my voice on this. It provided the perfect platform to invite more female innovators like me. Slowly I realized this divide existed because girls doubted their abilities. So I invited a few junior girls in my college, trained them in Arduino* programming, and helped them make their very first project. Eventually, we were able to maintain a 50-50 gender ratio at the lab.
When I first entered the workforce, I was fortunate enough to have a good mentor and advisor. I have seen many global companies trying to go for more women-centric recruitment drives. But there is still a decrease in the number of women across different job roles in a company. As a researcher, I find very few women who I can relate to in my field. In conditions like this, it is important to have a mentor who recognizes this problem. My mentor, Mr. Malaikannan Sankarasubbu, Vice President at Saama Technologies, has been an avid supporter of my work. He has pushed me to be a speaker at technical talks, financially supported my meetups, and has always believed in me. This has given me the confidence to pursue disrupting and novel research. He also helped me start the Chennai Chapter of Women in Machine Learning and Data Science (WiMLDS). I’ve been able to reach out to a greater number of women and increase their participation in STEM fields through WiMLDS Chennai.
What projects are you working on now?
Fortunately, my inclination for research translated into my first job, as a deep learning researcher at Saama Technologies. Saama Technologies helps pharma and biotech companies speed up their clinical trials, enabling them to bring drugs to market more quickly. It does so by using data optimization techniques in various parts of the pipeline for clinical trials. One such technique is machine learning. Saama has exposed me to a range of problems and projects. I have been able to use my wide areas of interdisciplinary knowledge in diverse projects. Quantum computing, computer vision, signal processing, and edge are just some of the technologies my projects have been based on at Saama Technologies.
For example, I have worked on optical character recognition (OCR) for recognizing handwriting in doctor’s prescription forms. OCR for handwritten text is one of the hardest problems in computer vision. We used OpenCV along with a convolutional neural network that estimates n-gram frequency (CNN-N-Grams) to recognize the text. This was the topic of my talk at this year’s PySangamam IIT Madras.
I have also worked on deploying machine learning to the edge. During this time, I realized that the Intel® Movidius™ Neural Compute Stick (NCS), a hardware accelerator, was marvelous in speeding up inference. This helped me run a face recognition model on it. Further, I gave a talk on “Running TensorFlow* Models on the Edge” at PyCon Malaysia 2018.
Tell us about a technology challenge you’ve had to overcome in a project.
India is one of the countries with the largest losses in power theft. We lose over $4 billion every year to power stolen in rural and urban areas. During December of 2016, the government of India started a Smart India Hackathon to address the various problems faced in the country. The Ministry of Steel, in particular, had a problem statement of detecting power theft via a mobile app. In India, urban areas have the problem of tapping power lines. Usually, the perpetrator clips the existing main line in a busy area redirecting power to his/her own house.
The Government wanted a way of registering this crime by taking a picture through a mobile application. But there was also a greater need for understanding which particular house was receiving the stolen power?. So, we implemented a machine learning algorithm along with the mobile app. Long Short-Term Memory (LSTM) was used to recognize anomalies in the time series data of power values for different households in an area. Furthermore, this acted as verification for the crime, which helped the electric power authorities. We won first prize for this solution and I published my first paper in an IEEE conference.
How are you planning to leverage Intel® technologies in your work?
The main bottleneck in any machine learning project is training. Intel has a lot of optimized tools and products that I use on a daily basis to decrease training time. I also feel that even though training takes time, we perform inference more often than we train models. In this respect, I use the Intel® Distribution of OpenVINO™ toolkit and the Intel Movidius NCS regularly in my work to optimize vision problems at the edge. Moreover, a lot of what I’ve learned about optimizing machine learning is from the awesome webinars that are conducted by Intel. I especially liked the speed up Python webinar and I regularly use a lot of what I learned from it in my daily work.
In addition to what I’ve been able to take away from the webinars that Intel has hosted, I consistently use Intel's hardware and software stack in my work. This technology has helped me meet client requirements, which have led to more recognition of my work. The Innovator program offers an amazing platform to reach worldwide developers and researchers, which has helped me tremendously in regard to my technology problems. As a woman in technology, it acts as encouragement and constantly bolsters my work.
What trends do you see happening in technology in the near future?
The advancement of technology is difficult to encapsulate through one’s individual efforts. However, the world of social media and collaborations via meetups helps us to expand our knowledge base. Platforms such as Twitter*, Hackster.io*, developer groups such as the Intel Software Innovator program, and Meetup* groups in my city have helped me stay updated with the latest technologies. One of the most fascinating technologies I have ever come across is quantum computing. Quantum mechanics has always been popular among physicists. And now with Dwave, IBM, and others offering quantum computers to the public, we are getting increasingly closer to a quantum computing-driven world. Although it is still a few years away, I see quantum computers being used for solving large optimization problems and other computationally expensive tasks.
Outside of technology, what type of hobbies do you enjoy?
I am a firm believer in giving back to the community. As a woman in technology, I am always working to bridge the divide. I am the co-organizer of women in machine learning and data science Chennai. I love organizing meetups and helping other women like me learn something new. This has also helped me teach women in the workforce new skills. Machine learning is a vital skill for women in data science and can add value to their work. I also help college juniors with their projects by sharing resources or connecting them to other women in my network.
Apart from that, I love reading and blogging. I come across many ideas as well as some bottlenecks in my projects, and I find that blogs are the best way to share your thoughts and experiences. Saama and open source communities like ‘IoT for All’ have been very supportive in publishing my work.
Learn more about the Intel® Software Innovator Program
Check out the latest innovator updates, get the program overview and learn about innovator benefits. For inspiration or to start sharing your work, join our community collaboration platform Developer Mesh.
Interested in more information? Contact Program Manager Wendy Boswell.
Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserverd for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.
Notice revision #20110804