ACADEMIC WORK
LIVE OPS PLAN : BALDUR'S GATE III
Time taken:
4 months
Tools used:
-
SteamDB
-
VG Insights
-
Kaggle Datasets
-
Google Sheets
-
Power BI
Description
As part of my Master's Course at ISART Business School, I was expected to create a LiveOps plan for an existing game.
The main aim of the project was to conduct through research and come up with content that has the ability to engage, retain and effectively monetize its player base and create room for new audiences.
I proposed accessibility improvements, an Epilogue DLC and a crafty Dungeon Master mode. These features are backed by data, community feedback and larger market trends.
LEVEL DESIGN : HAMMER
Time taken:
3 Weeks
Tools used:
-
Hammer
-
Half Life 2
Description
As part of a college assignment on level design we created a single-player level for Half-life 2 using the Hammer level editor. We were heavily inspired by the infamous Kowloon walled city that existed in the 90's. Our intentions were to create a level that included:
-
Claustrophobia with narrow spaces and clustered corners
-
Surprise in catching the players off guard
-
Spatial differences that allowed a varied gameplay
The following exercise allowed us to create an interesting playable level and enabled us to learn using a powerful tool for level design.
TOOL PROGRAMMING




Time taken:
7 Days
Tools used:
-
Unity
-
Editor Scripting
-
Property Drawers
Description
I have tried editor scripting earlier to make custom inspectors but this time I wanted to create something that has multiple use cases and increases productivity.
I got my idea from a designer who suggested that it would be great to have comments on game-objects when they are selected. This inspired me to create an object manager that does the following.
-
Create instances of objects in an organised manner
-
Adding comments to objects in scene view regarding the type and use of objects
-
Single manager to label and comment objects with a single list for organisation
MACHINE LEARNING




Time taken:
7 Days
Tools used:
-
Unity
-
Neural Networks
Description
Machine learning has always been an intriguing subject and I wanted to explore it in video games. A workshop at our college allowed us to create machine learning in Unity from scratch. Through the week we created a racing game with the AI learning the map over time and getting better to traverse through without colliding.
At the end of the week we were expected to add our own features to improve or enhance the simulation. I created a tool that allowed users to switch between a car that can be driven and the AI. This was further improved with pole position comparisons and a track of checkpoints cleared. The switch between the player and the AI was simply with a single button.
LEVEL DESIGN : PROBUILDER
Time taken:
14 Days
Tools used:
-
Unity
-
ProBuilder
-
ProGrids
Description
Working on the Probuilder for Unity, I managed to create a 3D level for a cycling game. The game is called Bubble Gum! and it was part of an individual assignment to create a level for the same.
I wanted to maximize the freedom of movement that the game offered. In order to achieve this I gathered inspiration from towns on slopes that allow players to explore verticality.
Here you can have a look at a gameplay montage of the level.