2R2T

2R2T is an emerging amateur esports platform aiming to become the reference for competitive League of Legends in France. The site was designed desktop-first with complex animated layouts and floating game assets, but had no mobile experience — cutting off a majority of users. Additionally, the platform needed a fair and transparent scoring system to evaluate players across different roles and skill levels, powering features like 'RateMyLvl' and team matchmaking.

Rebuilt the entire responsive layer from scratch, adapting the desktop-first design — with its radial gradients, floating 3D assets, and custom 'Phonk' typography — into a fluid mobile experience without sacrificing the brand's visual identity. On the algorithm side, co-developed the player rating engine: a weighted scoring system that evaluates players based on role-specific KPIs (KDA, vision score, CS/min, objective control) normalized across rank tiers, producing a single composite score used for matchmaking and leaderboards.

# Player rating algorithm — composite score
def compute_player_score(stats: PlayerStats) -> float:
    role_weights = ROLE_WEIGHTS[stats.role]

    raw = {
        "kda": stats.kills_assists / max(stats.deaths, 1),
        "cs_min": stats.cs / stats.game_duration,
        "vision": stats.vision_score / stats.game_duration,
        "objectives": stats.objective_control,
    }

    # Normalize against rank tier averages
    normalized = {
        k: (v - TIER_AVG[stats.tier][k]) / TIER_STD[stats.tier][k]
        for k, v in raw.items()
    }

    return sum(normalized[k] * role_weights[k]
               for k in normalized) * 100 + BASE_SCORE