The following is a basic outline of my SOP which I would be submitting for the upcoming fall intake. Please let me know what you think (Please be brutally honest :) )
During my stint as an undergraduate student at XXX Institute of Technology, I had the opportunity to contribute to the research on advancement of digital twins for electric vehicles undertaken by the university's Advanced EV Lab. We were trying to build generative time series models that could mimic the charging-discharging rates of Lithium-ion batteries under various stages of battery degradation. This data would eventually be used by the digital twin to predict the state of charge and state of health of the battery. I had come across contrastive learning methods while conducting a literature review and decided to experiment with them for this model. The battery charging-discharging data was mostly cyclical with a randomly varying frequency and my hunch proved right when my contrastive learning model managed to achieve an MSE of 0.3, significantly enhancing the accuracy of SoC predictions.
Additionally, while building this model, I got to use some experimental hardware developed by <redacted company name> Computing. <redacted company name> had set up FPGA based accelerators on campus as a part of the university's industrial collaboration program. Their processors were built on a customized and proprietary architecture that enforced parallel processing, which was able to offer processing speeds that I had never seen before - slashing the training and testing times by nearly half. I was fascinated by these processors and attempted several times to build a similar processor on a smaller scale (on an FPGA) in my personal time. However, I quickly discovered that designing such accelerator circuits was far more complex than I had anticipated. I lacked the knowledge of processor designs beyond 8-bit CPUs and despite my best efforts, I was unable to replicate even a simple vector processor. These failures built a strong desire to learn and explore computer hardware, particularly in the circuits that they are composed of.
At the end of my undergraduate studies, I had a keen liking towards machine learning and embedded systems. I wanted to build a career that would preferably involve both of them. Upon graduating, I managed to land a software engineering role at [YYY] and decided to further explore my interests in machine learning systems. [YYY] is an NLP firm that offers chatbot services for small-scale businesses in North America. At [YYY], I developed a BERT-based QA recommender system, with a precision and recall of over 0.9, capable of responding to both generic and situation-specific FAQs which was later patented. With the advent of ChatGPT in 2022, I was actively involved in integrating LLMs into [YYY]'s backend framework. I also implemented a custom Retrieval-Augmented Generation (RAG) pipeline aimed at sourcing information from multiple repositories and supplementing the existing QA recommender system with the latest relevant information. All my efforts led to the automation of over 600 customer interactions every day and the reduction of system response time by two seconds.
Having gained valuable experience in developing, implementing and maintaining production grade software systems over the last two years, I now wish to fulfill my desire to build my expertise in computer hardware and embedded systems. Having dealt with server crashes due to memory spikes, azure downtimes during peak hours and GPU overloads, I understand the significance of efficient and reliable hardware in the IT industry. I believe that by extending my expertise in both hardware and software, I can bring a new perspective to industries and be of more value.
With this goal in mind, I wish to pursue a master's degree in Electrical and Computer Engineering at [ABC] University. [ABC] offers me a chance to interact and collaborate with some brilliant minds making it an ideal place for me to fulfill my dreams.
After completing my master's, I plan to pursue a PhD in Electrical and Computer Engineering. I wish to pursue research in the field of processor architecture and embedded systems. With the ever increasing demand for consumer electronics, there is a global impetus to manufacture energy efficient hardware and this is where I would like to contribute - design energy efficient electronics for the world. In light of this, I hope to be part of the [OOO] Lab where I seek to contribute to the ongoing research on private computation accelerators. I believe that my previous work with hardware accelerators would make me an ideal fit for the lab. I am also passionate about developing autonomous systems, with a focus on implementing low latency machine learning solutions that operate closer to hardware.
In the long term, I aspire to establish my own company in the field of computer vision, focusing on advancing fully automated manufacturing. One sector I find particularly promising is the textile industry, which currently struggles with a reliance on seasonally available manual labor and faces ongoing shortages in its workforce. By introducing completely automated solutions, I hope to address these challenges and contribute to the industry's evolution. I wish to shape the future of this sector and drive innovation over the upcoming decades.
Another sector that has (unexpectedly) captured my attention in recent years is the financial industry, particularly the niche domain of High-Frequency Trading (HFT). This interest was sparked by conversations with a close friend who works as an FPGA engineer at a HFT firm. His endless stories about how cutting-edge hardware solutions can revolutionize trading systems has aroused my curiosity. I am intrigued by the potential of hardware acceleration and low-latency systems in this context and the competitive advantages that they have to offer in trading. I hope to dive deeper into this fast-paced industry and contribute to innovations in financial technology.
A few questions -
1. I have not mentioned my majors in undergrad - should I? I am under the assumption that adcoms will also take a look at my resume
2. Is it necessary to mention a professor's name in my sop (unless specified in the requirements)? I feel that if I use names, I would be reducing the number of people who actually take a look at my application
3. Should I mention a timeline ? - like graduated in 2022, started work in 2022 etc. Again, I am under the assumption that adcoms will also take a look at my resume
During my stint as an undergraduate student at XXX Institute of Technology, I had the opportunity to contribute to the research on advancement of digital twins for electric vehicles undertaken by the university's Advanced EV Lab. We were trying to build generative time series models that could mimic the charging-discharging rates of Lithium-ion batteries under various stages of battery degradation. This data would eventually be used by the digital twin to predict the state of charge and state of health of the battery. I had come across contrastive learning methods while conducting a literature review and decided to experiment with them for this model. The battery charging-discharging data was mostly cyclical with a randomly varying frequency and my hunch proved right when my contrastive learning model managed to achieve an MSE of 0.3, significantly enhancing the accuracy of SoC predictions.
Additionally, while building this model, I got to use some experimental hardware developed by <redacted company name> Computing. <redacted company name> had set up FPGA based accelerators on campus as a part of the university's industrial collaboration program. Their processors were built on a customized and proprietary architecture that enforced parallel processing, which was able to offer processing speeds that I had never seen before - slashing the training and testing times by nearly half. I was fascinated by these processors and attempted several times to build a similar processor on a smaller scale (on an FPGA) in my personal time. However, I quickly discovered that designing such accelerator circuits was far more complex than I had anticipated. I lacked the knowledge of processor designs beyond 8-bit CPUs and despite my best efforts, I was unable to replicate even a simple vector processor. These failures built a strong desire to learn and explore computer hardware, particularly in the circuits that they are composed of.
At the end of my undergraduate studies, I had a keen liking towards machine learning and embedded systems. I wanted to build a career that would preferably involve both of them. Upon graduating, I managed to land a software engineering role at [YYY] and decided to further explore my interests in machine learning systems. [YYY] is an NLP firm that offers chatbot services for small-scale businesses in North America. At [YYY], I developed a BERT-based QA recommender system, with a precision and recall of over 0.9, capable of responding to both generic and situation-specific FAQs which was later patented. With the advent of ChatGPT in 2022, I was actively involved in integrating LLMs into [YYY]'s backend framework. I also implemented a custom Retrieval-Augmented Generation (RAG) pipeline aimed at sourcing information from multiple repositories and supplementing the existing QA recommender system with the latest relevant information. All my efforts led to the automation of over 600 customer interactions every day and the reduction of system response time by two seconds.
Having gained valuable experience in developing, implementing and maintaining production grade software systems over the last two years, I now wish to fulfill my desire to build my expertise in computer hardware and embedded systems. Having dealt with server crashes due to memory spikes, azure downtimes during peak hours and GPU overloads, I understand the significance of efficient and reliable hardware in the IT industry. I believe that by extending my expertise in both hardware and software, I can bring a new perspective to industries and be of more value.
With this goal in mind, I wish to pursue a master's degree in Electrical and Computer Engineering at [ABC] University. [ABC] offers me a chance to interact and collaborate with some brilliant minds making it an ideal place for me to fulfill my dreams.
After completing my master's, I plan to pursue a PhD in Electrical and Computer Engineering. I wish to pursue research in the field of processor architecture and embedded systems. With the ever increasing demand for consumer electronics, there is a global impetus to manufacture energy efficient hardware and this is where I would like to contribute - design energy efficient electronics for the world. In light of this, I hope to be part of the [OOO] Lab where I seek to contribute to the ongoing research on private computation accelerators. I believe that my previous work with hardware accelerators would make me an ideal fit for the lab. I am also passionate about developing autonomous systems, with a focus on implementing low latency machine learning solutions that operate closer to hardware.
In the long term, I aspire to establish my own company in the field of computer vision, focusing on advancing fully automated manufacturing. One sector I find particularly promising is the textile industry, which currently struggles with a reliance on seasonally available manual labor and faces ongoing shortages in its workforce. By introducing completely automated solutions, I hope to address these challenges and contribute to the industry's evolution. I wish to shape the future of this sector and drive innovation over the upcoming decades.
Another sector that has (unexpectedly) captured my attention in recent years is the financial industry, particularly the niche domain of High-Frequency Trading (HFT). This interest was sparked by conversations with a close friend who works as an FPGA engineer at a HFT firm. His endless stories about how cutting-edge hardware solutions can revolutionize trading systems has aroused my curiosity. I am intrigued by the potential of hardware acceleration and low-latency systems in this context and the competitive advantages that they have to offer in trading. I hope to dive deeper into this fast-paced industry and contribute to innovations in financial technology.
A few questions -
1. I have not mentioned my majors in undergrad - should I? I am under the assumption that adcoms will also take a look at my resume
2. Is it necessary to mention a professor's name in my sop (unless specified in the requirements)? I feel that if I use names, I would be reducing the number of people who actually take a look at my application
3. Should I mention a timeline ? - like graduated in 2022, started work in 2022 etc. Again, I am under the assumption that adcoms will also take a look at my resume
