Nebula AI Core
A revolutionary dashboard transforming how we visualize complex financial data streams. The goal was to make high-frequency trading feel calm and intuitive.
Overview
Nebula AI Core is a real-time financial analytics dashboard built for quantitative trading desks. It ingests over 50,000 data points per second from global exchanges and surfaces actionable patterns through a spatial interface designed for low-latency decision-making.
Role
Led a cross-functional team of 8 — 3 frontend engineers, 2 backend, 1 data scientist, 1 designer, and 1 QA. Owned product strategy, sprint planning, and stakeholder alignment with the trading desk leadership.
Goals
Reduce average decision latency from 12 seconds to under 3 seconds. Achieve 99.97% uptime during market hours. Decrease onboarding time for new traders from 2 weeks to 3 days through intuitive spatial navigation.
Constraints
Strict regulatory requirements (MiFID II compliance) limited which data could be cached client-side. The existing legacy system had to remain operational during the 6-month migration window. Budget capped the team at 8 headcount.
Process
Started with a 3-week discovery sprint: shadowed 12 traders, mapped their workflows, and identified 7 critical pain points. Built a clickable prototype in Figma, ran 3 rounds of usability tests with real traders, then moved to a phased rollout starting with the options desk.
Key Decisions
Chose WebSocket streaming over polling for sub-second updates. Adopted a spatial z-axis layout instead of tabs, letting traders see correlated instruments simultaneously. Rejected a full microservices rewrite in favor of a strangler-fig migration pattern to reduce risk.
Results
Decision latency dropped to 2.1 seconds — an 82% improvement. System achieved 99.98% uptime in the first quarter. Trader satisfaction scores rose from 3.2 to 4.7 out of 5. The platform processed $2.3B in transactions during its first 90 days of operation.