Building an Agentic F1 Strategy Engine with InfluxDB 3
Session Date: Jul 28, 2026
Time: 8:00am (PT) | 3:00pm (GMT) | 4:00pm (BST)
Optimizing an F1 pit strategy means processing dozens of live telemetry signals—tire degradation, lap times, rival positioning—and making a call in seconds. Most data stacks weren’t designed for that velocity; queries get complicated, and pipelines get brittle.
Rohan Chandrashekar built a five-agent AI system, orchestrated with LangGraph, to predict and optimize F1 pit and tire strategies in real-time. InfluxDB is the data layer, handling live ingestion from F1 APIs, powering 32 dashboard panels with no pre-aggregation, and storing LLM-generated race debriefs alongside raw telemetry. Moving from InfluxDB 2.x to InfluxDB 3 simplified the hardest queries: positions-gained calculations that previously required three separate queries collapsed into a single SQL subtraction.
In this session, Rohan walks through the full architecture: how the agents are structured, how InfluxDB handles real-time ingestion and historical data side by side, and what changed when moving from 2.x to InfluxDB 3 SQL.
