Introduction
Sachin C -
Welcome to my Senior Project Blog!
Hi everyone! I’m excited to welcome you to my Senior Project Blog. This space will document my journey as I explore my chosen area of research during Trimester 3.
Project Title: Modeling the Energy Consumption of Compact Image Recognition AI Models on Microcontrollers
Thought it sounds complicated, I will break down complex jargon into simple-to-understand terms as we proceed.
About Me:
My name is Sachin Chandra, and I’m a senior with an intended major in computer engineering. Since the boom of OpenAI technology in November of 2022, I acquired an interest in developing artificial intelligence solutions to many of the problems we face in day-to-day life.
Project Overview:
Cloud computing, the delivery of computing services over the internet, was touted as a life-changing discovery in the early 2000s. Cloud computing offered faster innovation, flexible resources, and economies of scale that traditional computing could not. Delivering servers, storage, databases, networking, software, analytics, and intelligence over the internet (the ‘cloud’) was seen as revolutionary for a time period where hardware had to physically travel to consumers in order to be implemented.
Today, data security and privacy is at the top of the list of concerns for many, as widespread data breaches keep companies on their toes. In 2024 alone, over 10,600 data breaches occurred, costing users $4.88 million1, 2. For that reason, companies bear the burden of preventing identity theft, fraud, and other malicious activities by keeping their users’ data safe.
With the boom of artificial intelligence, cloud computing has become especially relevant to the discussion of data privacy and security as over 70% of AI services are delivered to companies over the cloud3. Because computing a personalized solution for a user requires a significant amount of data traffic, data privacy regulations are starting to take effect regarding users’ data. The most comprehensive of these regulations to date is the California Consumer Privacy Act (CCPA) and the California Privacy Rights Act (CPRA)4, requiring businesses in California to inform consumers how their personal information is collected, used, and retained, and protect personal data with reasonable security measures.
The issue lies in the scale of these regulations. Though they are a good step toward defining how data usage can remain between ethical lines, as of current they are only in place in California, nowhere near global enough to influence companies around the world to shift their focus more heavily toward data privacy.
Today, what defines ‘security’ of cloud-based AI solutions is still ambiguous. Therefore, many companies that handle delicate information are turning to native AI computing in order to protect especially sensitive or important data.
However, native AI computing has traditionally required a significant amount of extremely capable hardware and power, inaccessible to most due to their cost and high knowledge barrier. Therefore, finding a middle ground between economical hardware and energy-efficient software is necessary to make native AI accessible to more people.
Implications:
As AI becomes more integrated into our everyday lives, energy efficiency and privacy are growing concerns that many have. By finding solutions that can run on hardware as small and cheap as a microcontroller, we can reduce dependence on cloud services, lower latency, and improve security—all while making technology more accessible.
Thank you for joining me during my research. I’ll be sharing regular updates along the way. I’m excited to dive into this subject matter and report back!