Kunal Agarwal

×

PintOS (C)

Built integral parts of the PintOS operating system as part of university operating system class.

Wrote system calls including exec and wait, as well as important file system calls including write, read, and create.

Designed a file system from, modelled after UNIX Fast File System (FFS). Also, created a buffer cache to complement the file system.

Worked in a group of three others, and learned important skills on how to work on a large project in a group.

Learned how to navigate a large codebase and edit areas that need editing, based on how the codebase fits together.

(Code can be provided upon request)

×

Coursestamp (Flask)

Allows users to pick an online course on Youtube and search for all occurrences of a certain keyword or phrase in the online course.

Caption data from Youtube is placed into JSON files via a script. This data is then parsed through upon a request to find the video and the place in the video that the request is mentioned.

Uses Houndify voice to text software to implement an option for users to speak their request instead of typing it out.

Worked in a group of three others, and learned important skills on how to work on a large project in a group.

Created at TreeHacks 2020.

Devpost Code

×

Guardian (Android SDK/Pandas)

Uses the RideOS API to create an application that uses ridesharing technologies to match users with a designated 'guardian'.

The app chooses the safest path using imported crime data.

Used Berkeley Crime Data to import coordinates of latest criminal activity to pinpoint crime hotspots. Used Pandas to clean the data to put it into a format usable by the RideOS API.

These coordinates are inputted into the RideOS API, which uses a pathing algorithm to avoid these areas.

Created at CalHacks 2019. Won an award.

Devpost Code

×

Graph API and Trip-finder (Java)

Created a graph package using Java that allowed users to create a graph data structure for their own use.

Included in the API was a general traversal algorithm as well as breadth-first and depth-first traversal algorithms.

Also included in the API is an algorithm that determines the shortest path between two nodes in the graph. The algorithm uses A* and Dijkstra's to find the shortest path.

In addition to the graph API, I created a trip-finder that uses the shortest path feature of the API to find the best set of directions between two locations.

The example above shows the shortest path from Berkeley to Santa Cruz, with a waypoint in San Francisco.

(Code can be provided upon request)

×

Amazons: Artificial Intelligence (Java)

The game is called Amazons in which players move their queen and throw spears to trap their opponent.

I built the main functionality of movement in the game, the AI that the user could play against, as well as the GUI.

The AI utilized a minimax algorithm that was made more efficient with the use of alpha-beta pruning as well as iterative deepening. My AI placed in the top five of my data structures class of almost a thousand people. My AI can be seen playing itself above.

(Code can be provided upon request)

×

Personal Website and Teaching Materials (Django)

Created a web application that allows for the searching of worksheets based on the topic desired.

The web application was created using Django, allowing for quick addition and look-up to the SQL database storing all the worksheets.

I also created and designed a number of worksheets for CS 61A, an introductory computer programming class here at UC Berkeley. This has allowed me to achieve an expert level understanding of the nuances of python.

×

Scheme Interpreter (Python)

I used mutual recursion to evaluate scheme expressions, and then apply the given expression. I was able to do this using Python.

In addition, I ensured no extra frames were created if the expression was in a tail context, allowing for tail recursion.

(Code can be provided upon request)