Rongbiao Gu

Analytics and agent harness engineering

• Self-initiated and built solo a self-serve analytics and knowledge platform with ~120 active users/month (~30% of company) across PM, CSM, engineering, design, and GTM, spanning SQL Q&A, dashboard and HTML report generation, product-truth Q&A, and codebase/change-history queries. Supports 1,000+ conversations/month and dozens of user-authored context artifacts and skills. Achieved this adoption despite company-wide access to Claude Code/Cowork.
• Architected the underlying agent harness with a live documentation pipeline (code + Confluence) into a tiered, human-verified knowledge index.
• Built an agent-as-judge eval framework to mine conversations for failure modes — tool-call performance, knowledge gaps, behavioral gaps like missed clarifying questions or weak citation/data-lineage. Then used a stronger model to propose fixes, with everything gated through human review before it shipped.

• Built data pipeline and credit risk models for 0-to-1 credit card product launch.
• Worked on data governance regime in the strictly regulated financial environment, including data safty and privacy protection

Hardware decision - bikes, scooters and docking stations

• Architected and deployed end-to-end cloud infrastructure (GCP) and real-time computer vision models for IoT preventative maintenance.
• Built computer vision platforms for industrial and transportation edge applications.

• Built an e-commerce asset platform that automates discovery, outreach, and project management for the aquisition team, it supported over $10M in brand acquisitions.
• Built end-to-end automated workflow tools for brand managers to streamline operational, from listing management, competitive intelligence to ad spend.

• Revamped core ad-bidding algorithm architecture in PyTorch to be marketplace-agnostic across multi-channel e-commerce platforms.
• Built deep learning encoder models and an automated advertising keyword recommendation system.

• Co-founded and built a SaaS platform delivering default risk predictions across millions of Canadian consumer loans.
• Managed product roadmap, data pipeline engineering, and customer integrations end-to-end.

• Engineered core bidding systems: designed the first real time bidding algorithm that 60x the legacy algorithm
• Built workflow automation application for programmatic ad campaign optimization. This product scaled our operation by 4x without adding headcount

• Executed M&A technical and data due diligence for target acquisitions in the media and healthcare sectors.
• Developed data strategy roadmaps and pre-sales technical assessments to evaluate target company viability.