Eshika Pathak

Hi, welcome to my corner of the internet!

I'm an Electrical and Computer Engineering graduate student specializing in Machine Learning at Carnegie Mellon University .

My engineering journey began at IIT Gandhinagar, India, and since then, I've immersed myself in hands-on projects across autonomous systems, safe control, optimization, and reinforcement learning.

I believe that with time and practice, anything can be learned—nothing is innate. My goal is to design intelligent systems that learn and adapt in the same way. There's nothing I enjoy more than learning, experimenting, and exploring new ideas—except maybe playing devil's advocate!

Email / CV / LinkedIn

I read some papers outside my projects and write my thoughts on them. Here are my [notes]. Let's discuss!

This page is still under construction!

Projects

Autonomous Vehicle Safe Control in the Presence of Self-Seeking Humans
Research Assistant, Control and Learning Group, Carnegie Mellon University (August 2023 - Present)
Designed and experimentally validated optimal and safe autonomous vehicle policies by accounting for humans’ self-seeking behaviors in a game-theoretic setting.

Adaptive Risk-Aware Reinforcement Learning Based Multi-Stock Portfolio Optimization [presentation]
2024
Integrated deep reinforcement learning agents with a control barrier function-based controller for optimal risk control.

Personalized Federated Learning using Hypernetworks [report]
2024
Reproduced and performed ablation study with insights on client learning configurations and hypernetwork architectures.

Information Coding and Learning in Spiking Neural Networks and Predictive Coding Networks [report]
2024
Applied SuperSpike and differential equation-based learning for image classification in SNNs and PCNs.

Blind Source Separation in High-Density EMG Data [report]
2023
Designed algorithms for noise filtering, wavelet feature extraction, PCA, NMF, and clustering to analyze motor unit signals.

Bidirectional Charger Models, V2X Framework & Optimal Decision Making
Research Intern, Bosch Center for Artificial Intelligence (May 2022 - July 2022)
Developed two bidirectional charger models for a consumer electric vehicle. Built an end-to-end simulation framework for vehicle-to-grid and vehicle-to-house (V2X) power transfer capabilities and optimal multi-variable decision-making for profit.

OPC Algorithms in Computational Lithography [presentation]
Research Intern, Nano Devices and Circuits Lab, IIT Gandhinagar (August 2021 - May 2022)
Built a robust resist and optical model lithography simulation system with control on critical optical proximity correction (OPC) parameters. Formulated an intelligent mask fragmentation algorithm to reduce computation time-accuracy trade-off.

Reinforcement Learning Algorithms for Autonomous Navigation [report]
Research Intern, Learning and Emerging Networked Systems Lab, Texas A&M University (May 2021 - July 2021)
Modeled deep reinforcement learning algorithms for autonomous navigation of mobile robots in indoor environments. Engineered and tested 11 reward functions. Trained and deployed an optimal model demonstrating perfect waypoint tracking.

Automating Microbial Growth Detection and Monitoring Using TDLAS [presentation]
2023
Detected E.coli growth with quantum cascade laser and TDLAS; designed an intelligent algorithm to automate baseline selection.

Wearable Foot Plantar Pressure Monitoring and Analysis System [presentation]
2022
Engineered a system to analyze gait data from shoe insole sensors, reducing latency from over a minute to seconds.

Energy Efficient Control of An Inverted Pendulum [report]
2022

Mathematical Modeling and Analysis of COVID-19 Pandemic Using SEIR [presentation]
2021
Devised an extended Susceptible-Exposed-Infectious-Removed (SEIR) numerical model to account for vaccination, deaths, and limited beds and predict the future of the COVID-19 pandemic. Analyzed Indian government data using the model.

Teaching

TA for 18662: Principles and Engineering Applications of AI
Carnegie Mellon University (Spring 2024)
Assisted in delivering lectures, preparing course materials, homeworks, and guiding graduate students through complex machine learning concepts via examples. Some of the material I made and delivered:

PyTorch, MLPs, CNNs [python]
Expectation Maximization, Logistic Regression [slides] [python]
Stat ML, Bayesian Updating [slides] [python]
Conditional Independence, MiniMax for Optimal Decisions [slides] [python]
Dijkstra, Bellman Ford Algorithms [slides] [python]
Probability Theory Review [slides]

TA for MS403: Engineering Entrepreneurship
IIT Gandhinagar (Fall 2022)
Facilitated discussions, graded assignments and exams, and supported students in understanding applications, brainstorming and projects.

That's about it

Interesting people to read about: Diogenes, Seneca, R.D. Laing .

It's only fair that I get to know you too. Shoot me an email (epathak@andrew.cmu.edu) and let's chat!

Website template credit to Jon Barron.