Nat Lu

Connecting models, data, and product into one system

Nat leads the integration of AI pipelines at KLAR, turning individual components into a working, end-to-end system.

With experience across data workflows and applied AI systems, she focuses on how models move from isolated outputs into something usable within a product. Her work sits at the intersection of data pipelines, agent workflows, and system integration, ensuring everything connects cleanly from input to output.

At KLAR, she builds and connects the pipelines behind KLARva, from data ingestion and scraper APIs to analyzer agents and the final recommender system. Working closely with the system design, she ensures that each component flows into the next, so outputs remain structured, consistent, and usable.

She also sets up the analytics layer, configuring clickstream tracking and dashboards to surface how users interact with the product. This allows the team to continuously refine the system based on real use, not assumptions.

Nat’s approach is grounded in making AI systems actually usable. Not just functional in parts, but coherent as a whole. She focuses on turning fragmented processes into connected workflows that support real product behaviour.

Her work at KLAR focuses on:

• Building and connecting end-to-end AI pipelines across agents
• Integrating data flows from scraper APIs to recommender outputs
• Ensuring consistency across analyzer and recommendation systems
• Setting up product analytics and user behaviour tracking
• Automating CI/CD pipelines to enable fast, reliable deployment