Hacker

KADHIR K

@kadhir

Joined Oct 2026Student at Indian institute of technologyIndia

Projects

  • VALOGRAPH — Real-Time Tactical Graph Intelligence & Persistent Agent Memory Engine for Valorant

    Graph HacksOct 2026

    What problem it solves: Competitive Valorant players, IGLs (In-Game Leaders), and esports analysts rely on fragmented stat trackers and external wikis that only show isolated flat tables (raw K/D or static weapon stats). Standard LLM assistants fail at live tactical gaming coaching because they hallucinate patch stats, cannot perform multi-hop relational reasoning across map geometry, agent skill synergies, player weapon mastery, and 10-player match rosters simultaneously, and forget a player's context once a session ends. Furthermore, jumping out to external browser tabs mid-game breaks focus. Who it is for: Competitive Valorant players, esports rosters, coaches, and solo-queue grinders who need an all-in-one, zero-redirect tactical command center. How VALOGRAPH solves it: VALOGRAPH is a real-time tactical graph intelligence platform powered by FalkorDB Cloud (`hacking_gg_valorant`) and Groq. Running locally alongside the game via `python app.py`, it renders an interactive 3D/2D node graph directly inside a custom dark-themed tactical UI—displaying all Agents, live Skill previews, Weapon damage profiles, Maps, and 10-player Live Squad / Recent Match rosters without ever linking out to external web pages. Every click executes live Cypher queries (visible in the bottom telemetry bar) to expand connected nodes, while the "AI Team Combo" and "Ask Falkor" agents traverse the player's subgraph and persistent `[:REMEMBERS]` nodes to deliver instant, personalized mid-match coaching.