Marvel's Avengers
Core project contributions...
Level Designer I | Crystal Dynamics | Horizon 2
-
Live Service Systems: Expanded a hex-based level assembly system into a live-service mission production pipeline, defining connection rules for level borders, elevation variation, and environment theming to enable rapid construction of hundreds of unique missions
-
Mission & Combat Design: Designed player objectives and mission layouts (defense, destruction, puzzle-gated vaults) for live-service content, and tuned enemy composition and encounter pacing as new enemy types entered development
-
Traversal Design: Authored traversal paths for all playable heroes, tailoring level geometry to each character's unique ability set — wall-cling, swing, flight, double-jump, wall-run
-
Key Responsibilities: Live Service Content, Visual Scripting, Combat Encounter Design, Mission Design, Traversal Design
Marvel's Avengers Level Design
In order to build the levels in Marvel's Avengers, we set priorities to guide our approach within our limitations. These goals were:
-
Build world maps for the live-service model
-
Create memorable locations players would travel to
-
Design levels and activities that support 1-4 players
Although there are lots of details that went into building these levels, here's how each of these goals shaped my approach:
Live Service
While DLC content like the raid was built as unique, hand-crafted spaces, I built our live-service multiplayer missions entirely from modular hexes — a mission-rotation model similar in spirit to games like Monster Hunter: World. I maintained a large enough library of hex pieces to assemble new map variations quickly, define how many objectives a given mission needed, and get it published on a weekly rotation without a long lead time. Enemy types, reward placement, and resource/hazard spawn frequency could also be adjusted after launch — though that specific tuning was driven by our live-ops team monitoring player data, not something I tracked myself. My contribution was building the system fast and flexible enough that those live adjustments were possible in the first place.
Level Variety
Variety came from two directions:
-
Biome: Exterior hexes spanned multiple themes — city, forest, dry canyon, and tundra — each accounting for height variation and border rules where hexes connected.
-
Faction presence: Interior hexes were AIM-themed, mostly robotic environments tied to the game's antagonist faction, accessible from exterior spaces through warp points. AIM's presence could be dialed from barely-there to a full base, giving missions distinct identities beyond just their exterior look.
Together, these two axes meant the same modular hex toolkit could produce missions that felt meaningfully different from each other, not just visually but in tone and threat level.
Player Experience (1-4)
A main focus of mine was designing mission spaces to support up to 4 players at once, each potentially playing a different hero with a different ability set. I designed around one core rule: every hero needed a viable path through the same content, without a single "golden route" that favored one kit over the rest. That challenge showed up across three layers:
-
Traversal: Every hero — wall-clinging, swinging, flight, double-jump, or wall-running — needed several viable routes from point A to point B, not just one. This was harder than it sounds inside a modular hex system, where geometry had to stay flexible enough to tile and reconnect while still offering distinct, interesting paths per ability.
-
Combat: Enemy types played easier or harder depending on the hero facing them. Flying enemies, for example, were more manageable for ranged heroes like Hawkeye and Black Widow, whose kits centered on rapid-fire projectiles.
-
Objectives: Interactables and mission objectives varied in difficulty depending on which hero attempted them, adding another layer of hero-dependent variety beyond movement and combat.
Together, these three layers meant a 4-player squad would experience the same mission differently depending on their hero picks — without any one hero's kit going unused or making the level trivial.
Level/Combat Layout Examples
- player ceiling
- player pathing
City traversal: Being creative with our traversal prims, we used pole-like structures as swing prims, along with rooftops, vehicles, and suspended scaffolding as platforms. Our visual language (VL) for traversable surfaces included flat surfaces, yellow tarps over structures, and white-painted handrails — visual cues that signaled to players where they could climb or interact.
Cover pieces: These were built mainly for enemy use — ranging from vehicles and plant dividers to barricades, HVAC units, and chimneys on rooftops.
Rewards | Resources placements: Locations varied between primary player paths and off-the-beaten-path spaces, to motivate exploration.
Health | Hazard placements: Health pickups were typically placed in low-traffic, safer areas, while hazards were placed in high-traffic zones.
Objectives: Objective space requirements varied in size and shape, sometimes fitting into tighter footprints depending on hero abilities. Enemy spawns near objectives were also tuned to expose specific hero vulnerabilities or strengths, adding another layer of hero-dependent difficulty — consistent with the combat variety described above.
- wall run/cling/scramble options
- player pathing
Hex Size: An Early Decision We Couldn't Walk Back
Hex size was one of the earliest decisions made on the project — sized to comfortably fit an objective's geometry, like a small enemy base, when the game only supported 2 players. That decision locked in before missions scaled up to 4 players, and by the time they did, it couldn't be revisited without reworking the entire hex system. At 4 players, with up to 16 enemies plus objective interactables all competing for space in the same hex, things got crowded fast. We made it work through careful combat and objective placement — but it's a real example of how an early technical decision can create constraints you're still solving around much later in production.
Designing Traversal for a Roster Still in Motion
Hero traversal abilities weren't locked in when I started building maps — they arrived in waves, and each one changed what "getting around the world" meant. Captain America came first, designed as our "Ryu" — the baseline, most balanced character everything else would be measured against. Iron Man came next, bringing flight into the mix; to keep pace, Cap was given wall-run and double-jump so he wouldn't fall behind traversal-wise. When Black Widow and Ms. Marvel were added with swing abilities, we realized they couldn't reach some spaces as easily as the rest of the cast, so we added a swing boost to close that gap. Hulk came last, with wall-cling and a jump that went higher the more consistently a player timed it.
Every new hero meant reassessing which traversal prims I needed, and how those prims communicated their use — "swing here," "wall-run there" — consistently across multiple themes and both interior and exterior spaces. I couldn't design once and move on; I was designing for a roster that kept changing underneath me.
Resource Design
As a design rule at Crystal Dynamics, we wanted players to constantly find and be rewarded for exploring — filling an otherwise empty world with things worth discovering. Using our custom map tooling, I could select any hex and set it to spawn resources — crafting materials, health pickups, hazards, or chests — chosen at random from a pool of pre-determined locations rather than fixed spots. That randomization mattered: without it, players could memorize spawn points and spawn-camp resources instead of actually exploring for them. Because this system was live-tunable, we could react quickly to real player behavior — for example, when chest spawns were too generous, we corrected the frequency live rather than waiting on a full patch cycle.