Projects
Projects
Under Professor Shrideep Pallickara, Colorado State University
Contributing to projects centered on supporting sustainability efforts by building GIS data exploration and analysis tools which utilize Big Data and Distributed Systems approaches.
Course project for Distributed Systems graduate course with Professor Shrideep Pallickara
A distributed, failure-resilient file system using replication implemented in Java and leveraging socket programming and multithreaded instances.
3 types of nodes were developed: a client, a chunk server, and a controller, running on a minimum of 10 distinct machines.
Files are uploaded and downloaded using a client. The controller manage information about the file chunks, including overseeing replication. The chunk servers manage the file chunks, including checksums.
Course project for Distributed Systems graduate course with Professor Shrideep Pallickara
Implementation of a distributed hash table (DHT) according to the Chord methodology.
Node entry is simulated by an extra discovery node, whose only task is to be aware of and return the address of a random node in the system (instead of through a broadcast protocol).
Node IDs are assigned non-deterministically using the same ID-space as the data items being stored. Every node is responsible for data items mapped between themself and the next-lowest node.
Unexpected node exits and failures are handled and routing tables repaired.
Course project for Distributed Systems graduate course with Professor Shrideep Pallickara
Nodes are set in a ring topology and during every 'round', determine some random amount of work to perform (calculating a nonce). Load is balanced among nodes in a centralized manner. Oscillating load transfers are avoided and final work done is verified with work generated to validate no lost messages.
Course project for Distributed Systems graduate course with Professor Shrideep Pallickara
The first assignment was geared toward setting up reusable communication mechanism in an asynchronous and multi-threaded distributed system.
A random topology, with weights, is generated by a centralized node. Dijkstra’s algorithm is leveraged to determine routing of basic packets within the system.
Under Dr. Nikhil Krishnaswamy, Colorado State University
Project developing an AI assistant for disorienting flight.
Developed an AI assistant for pilots addressing spatial disorientation challenges.
Designed and implemented an inverted pendulum simulation environment using Gymnasium to mimic aircraft-leveling scenarios.
Conducted extensive model training within the simulation environment, employing various machine learning techniques.
Collaborated in a multidisciplinary team, integrating AI models as assistants alongside a machine learning model that represented human participants.
Contributed to research in the field, co-authoring a paper published at HAI 2024.
Under Dr. Nikhil Krishnaswamy, Colorado State University
Project developing an AI assistant for middle school teachers with groupwork.
Developed multiple pipelines for collecting and processing user data. Pipelines included automatic speech recognition, data collection GUI (back- and front-end, currently in use at four data collection sites across the country), stereo vision.
Established calibration protocol for stereo vision using Microsoft Kinect SDK and OpenCV.
Led data collection.
Published two peer-reviewed research papers.