Hi!

In my freetime, I'm big into reading, muay thai, hiking, spending time with my dogs, cooking, and occasionally even writing some code ;)

Click to flip me!

Hirsh Guha

Hi! I'm a Software Engineer II at Amazon where I work in low-latency computing pricing ads, and defining fee structures throughout Amazon's DSP. I graduated from the University of Kansas School of Engineering with my degree in Computer Science in two years(first one ever!) by taking 34 credit hours a semester. My capstone project was building an online voting system.

I then went on to do my Master's and PhD. in Computer Science at Princeton under Dr. Andrew Appel, who promptly convinced me that secure online voting is infeasible. Lol.

Scroll to learn more about me!

Get in Touch

Contact Us

Want to contact me? Open the envelope!

My Research

Here are some of the research projects I've been involved in.

Password Pattern Analysis and Policy-Based Mask Generation For Behavioral Trends

Large scale corporate data-leaks happen almost weekly. Often, the most worrying leaked piece of information are passwords. Subsequently, penetration testing and security analysis companies make up a multi-million dollar industry, and the study of password security and password choosing behaviours co...

Link To Paper

DDoS Detective: A Collaborative Telemetry SystemFor DDoS Identification and Mitigation

Distributed Denial-of-Service (DDoS) is a broad class of network attacks that overwhelm a server with a high volume of traffic. This paper presents a system called "DDoS Detective" for cooperation between telemetry systems across ASes to detect and mitigate amplification based DDoS attacks that use ...

Link To Proposal Link To Paper

Safety from Bigoted AI: A Framework for Online Debiasing

Deployed models often stream high-impact outputs and actively train on newobserved data. Many fairness algorithms provide extremely strong guaranteesabout statistical parity on protected groups in the batchwise setting or in thelimit of training data. We are therefore motivated to extend these resul...

Link To Proposal Link To Paper

Face To Face: An Accuracy-Time Weighted Comparison of Facial Recognition Methods

Facial recognition has broad applicability in a variety of fields, including user authentication, person identification, video surveillance, criminal investigations, data privacy, gaming, and photography. In this paper, we study three different methods for categorizing faces from a variety of images...

Link To Paper

Classifying Twitter Sentiments from the Black Lives Matter Movement

We classify tweets related to the Black Lives Matter movement as positive or negative in order to ascertain an understanding of various classifiers. After cleaning the data, we train 5 different classifiers, each using both a bag-of-words representation and a tf-idf score. We use a variety of metric...

Link To Paper

Predictive Markers of Engineering Graduate Admissions Behaviour

In the area of Engineering Schools, some would describe the admissions process of Masters and PhD programs as something that would require a crystal ball to get right. Well that most certainly might help, there is evidence that Engineering Departments follow changes that can be predicated, and impro...

Link To Proposal

Password Pattern Analysis and Policy-Based Mask Generation For Behavioral Trends

Link To Paper

DDoS Detective: A Collaborative Telemetry SystemFor DDoS Identification and Mitigation

Link To Proposal Link To Paper

Safety from Bigoted AI: A Framework for Online Debiasing

Link To Proposal Link To Paper

Face To Face: An Accuracy-Time Weighted Comparison of Facial Recognition Methods

Link To Paper

Classifying Twitter Sentiments from the Black Lives Matter Movement

Link To Paper

Predictive Markers of Engineering Graduate Admissions Behaviour

Link To Proposal

Hirsh Guha

Get in Touch

Contact Us

Hirsh Guha

Get in Touch

Contact Us

My Projects

My Blog

My favorite books of 2024(So far!)

My Research

Password Pattern Analysis and Policy-Based Mask Generation For Behavioral Trends

DDoS Detective: A Collaborative Telemetry SystemFor DDoS Identification and Mitigation

Safety from Bigoted AI: A Framework for Online Debiasing

Face To Face: An Accuracy-Time Weighted Comparison of Facial Recognition Methods

Classifying Twitter Sentiments from the Black Lives Matter Movement

Predictive Markers of Engineering Graduate Admissions Behaviour

Password Pattern Analysis and Policy-Based Mask Generation For Behavioral Trends

DDoS Detective: A Collaborative Telemetry SystemFor DDoS Identification and Mitigation

Safety from Bigoted AI: A Framework for Online Debiasing

Face To Face: An Accuracy-Time Weighted Comparison of Facial Recognition Methods

Classifying Twitter Sentiments from the Black Lives Matter Movement

Predictive Markers of Engineering Graduate Admissions Behaviour

My favorite books of 2024(So far!)