Hive Dive
An endless-running on-rail shooter where you pilot a remote-controlled drone into hordes of enemies. Shoot down the invaders, fly through colorful Patagonia, and get the highest score!
Enemy design

For each enemy, I documented a full design package covering its purpose, visuals, and behavior.
Each color-coded element below corresponds to one part of my enemy design documentation.
-
Enemy Name – A clear and descriptive title.
-
Purpose – Why the enemy should be included in the game and what gameplay role it serves.
-
Behavior Sketches – Diagrams showing attack logic or area control patterns.
-
Visual References – Inspiration images for artists to ensure consistency.
-
Pathing Sketches – Visualizations of how the enemy moves through the level or interacts with splines.
Responsibilities:
-
Created enemy archetypes for distinct gameplay roles.
-
Documented purpose, behavior, and pathing for each type.
-
Balanced stats and telegraphs for fairness and clarity.

Encounter design
The encounters in this project were carefully crafted to keep players engaged and on edge,
with minimal downtime:
-
Players are given a maximum break of 2 seconds between encounters to maintain consistent action.
-
After turning corners, a 2-second break is provided to allow players to orient themselves
and anticipate the upcoming level layout.
From playtesting, I discovered key insights that shaped the design:
-
Longer breaks (exceeding 2 seconds) disrupted the pacing, causing players to feel
disengaged due to a lack of on-screen action. -
Simultaneous turns and enemy encounters proved overwhelming, as enemies fixed to the camera
created chaotic and unpredictable movement patterns.
These findings ensured a balanced and engaging player experience by addressing
pacing and encounter clarity.
Responsibilities:
-
Designed and paced enemy waves for continuous flow.
-
Controlled intensity with 2-second downtime limits.
-
Built challenge through enemy mix and positioning.
-
Tuned encounters for visibility and timing around corners.
-
Iterated after playtests to refine pacing and tension.
Level fly trough
Level design
This level was crafted using 20 tiles selected from 8 unique designs created by a visual artist on our team. To ensure smooth gameplay, splines were carefully placed at the correct heights, preventing players from flying into the ground or encountering sharp turns that could confuse them. The tile set includes 2 turns, 2 plains, 2 trees and 2 waterfalls.
The level features an endless runner format, where players navigate a circular path. With each completed lap, enemies grow progressively stronger, increasing their damage and health to add an engaging layer of challenge.
Responsibilities
-
Designed layouts with clear pacing and smooth drone flow.
-
Balanced visibility and navigation after sharp turns.
-
Iterated layouts to prevent clipping and confusion at high speed.

Light blue: Player spline Dark blue: Player start
Balancing (Player and Enemies)
Balancing was achieved through extensive playtesting, with the Sniper enemy posing the greatest challenge.
Its high-speed bullets made it a significant threat, even though it could be defeated with a single shot.
Instead of nerfing the Sniper, adjustments were made to empower the player:
-
The player’s fire rate was significantly increased, allowing for quicker attacks.
-
The bullet speed was also enhanced, making aiming more intuitive and effective.
These changes ensured a fair yet challenging experience, maintaining the enemy's threat level while
improving player agency and responsiveness.
Responsibilities:
-
Tuned stats for smooth difficulty progression.
-
Adjusted encounter pacing and time-to-kill ratios.
-
Collected and implemented playtest feedback.
-
Kept combat challenging but readable.
AI Implementation
Enemy behavior was implemented using a state tree, providing a simpler and more flexible alternative to complex behavior trees for prototyping and iteration.
Movement System:
-
All enemies, except turrets, use splines for navigation.
-
When an enemy reaches the end of its spline, it calculates whether it is to the left or right of the player's spline center and follows a corresponding spline to return to the center of the screen.
Shooting System:
-
Due to the player's fast movement, enemies often missed their shots when aiming
directly at the player’s current position. -
To address this, a predictive targeting system was developed. Enemies calculate the player’s
future position along the spline based on:-
Bullet speed
-
Player speed
-
Player offset from the spline center
-
This predictive system significantly improved enemy accuracy and ensured a challenging yet fair experience for players.
Predictive targeting system blueprint
State Tree Example:
Here is the state tree used for the enemies in this project. It outlines the flow of behaviors from movement along the spline to shooting actions. Each state triggers specific animations, movement, and attack behaviors, ensuring smooth transitions and responsive AI.
Responsibilities:
-
Set up simple state machines for enemy behavior.
-
Added predictive aiming using player velocity and bullet speed.
-
Tuned accuracy and delay for fair difficulty.
-
Adjusted movement along spline paths.
-
Fixed loops and synced logic with design goals.


Enemy Data Asset
State Tree
Turret shooting task
UI
Before making the settings screen, I looked at other games to see what options players usually have and how they interact with them. I then built all the UI logic for the settings screen and made sure it works smoothly with a controller.
Responsibilities:
-
Made all the UI logic for the settings screen
-
Got the UI working smoothly with a controller
-
Checked out other games to see what options players usually get
-
Made navigation simple so players can adjust settings without problems
What I Learned
-
Small design tweaks can completely change pacing and player flow, especially when testing enemy spawn timings
-
I started seeing encounters as part of a bigger rhythm, not just single fights, after playing through levels and adjusting enemy setups
-
I got better at tuning enemy behavior and readability while tweaking movement patterns, attack timing, and color feedback during playtests
-
Balancing goes way faster when you test early instead of waiting until everything feels finished
-
Small changes to movement speed or spacing often make more impact than adding new features
-
I learned how important it is to make UI work smoothly with a controller and be easy to navigate







