← Back

Business Intelligence Engineer II · AWS Marketplace

Building scalable decision systems for AWS Marketplace

5,000+ users · 11,000+ views · 40 hrs/week saved · 100M+ rows · Load time: 90s → 5s

Problem

Partner management teams were operating across fragmented, slow, and inconsistent data systems. Compliance tracking was manual. Revenue dashboards took 90+ seconds to load. Renewal metrics had no single source of truth. Different teams were working from different numbers and making conflicting decisions.

The cost wasn't just inefficiency. It was eroding trust in data across the org.

Goal

Build unified, high-performance decision systems that teams actually use: fast enough to be practical, accurate enough to be trusted, and designed to scale.

Approach

I treated each dashboard as a product, not a report. That meant starting with user research (conducting 10+ stakeholder interviews across 6 personas), writing Business Requirements Documents, and designing for adoption from day one. Performance and trust were product features, not backend concerns.

Driving Adoption

Building great systems isn't enough. Adoption requires intentional strategy.

  • User research upfront: Conducted 10+ stakeholder interviews across 6 personas before building. Ran group testing sessions with stakeholder groups to validate designs and iterate on feedback.
  • Launch emails with clear value: Sent targeted launch emails to relevant stakeholder groups explaining what the product does, why it matters, and how to use it.
  • "How to connect" embedded in every product: Added direct feedback channels in every dashboard so users could report issues, request features, or ask questions. This created a feedback loop that drove continuous improvement.
  • Feature enhancements based on feedback: Shipped 8 feature enhancements for the compliance dashboard and 9 for the revenue dashboard based on user requests. Adoption grows when users see their feedback implemented.
  • Performance as a growth lever: Reducing load time from 90 seconds to 5 seconds wasn't just a technical win. It was a product decision that directly drove activation and retention.

Result: 5,000+ users, 11,000+ views in 30 days for the compliance dashboard, 1,000+ views in 7 days for the revenue dashboard.

Key Projects

Partner Product Deployment Dashboard

Led end-to-end development of a compliance tracking solution for partner management teams. Delivered in 2 days by consolidating 4 disparate data sources.

  • 5,000+ users across partner management organizations
  • 11,000+ views within 30 days of launch
  • Reduced manual compliance tracking by ~40 hours/week
  • 8 feature enhancements shipped post-launch based on user feedback

Revenue & Insights Dashboard

Consolidated two legacy dashboards into a unified solution processing 100M+ rows. Engineered significant performance optimization as a core product goal.

  • Load time reduced from 90+ seconds to under 5 seconds
  • 300+ users across sales, partner management, and finance stakeholders
  • 1,000+ views within 7 days of launch
  • 9 major feature enhancements delivered based on stakeholder requests
  • Eliminated the need to maintain multiple dashboards

Renewals & Retention — Single Source of Truth

Led the definition and implementation of renewals tracking from ambiguous requirements. Architected renewal logic from scratch, incorporating multiple business scenarios serving both partner and customer teams.

  • Delivered single source of truth for renewals metrics in under 2 weeks
  • 180+ users across sales operations, finance, and partner management stakeholders
  • Eliminated inconsistent reporting across business teams
  • 600+ views within 6 months of launch

Product Decisions

Treated performance as a core product feature

Slow dashboards don't get used. I made performance a primary design goal from day one, not a post-launch optimization. Reducing load time from 90 seconds to 5 seconds wasn't a technical win. It was a product decision that drove adoption.

Tradeoff: Required deeper system optimization early, which added upfront complexity. The payoff was measurably higher adoption at launch.

Established a single source of truth despite stakeholder resistance

Multiple teams were attached to their existing renewal reports and metric definitions. I drove cross-functional alignment by facilitating deep-dive sessions across sales operations, finance, and data engineering to standardize definitions and consolidate into one system.

Tradeoff: Required significant stakeholder investment upfront. The result was a system teams trusted enough to replace their spreadsheets.

Designed for scalability, not just the immediate use case

The Row Level Security solution I built was the first of its kind for the organization. It was designed with enough flexibility that it became the standard data access control solution for 150+ dashboards across 12+ user personas, well beyond its original scope.

Tradeoff: More complex initial design. But it meant zero rework when the solution scaled org-wide.

Impact

  • 5,000+ active users across partner management and sales organizations
  • 40 hours/week saved in manual reporting across teams
  • Dashboard load time reduced from 90+ seconds to under 5 seconds
  • Single source of truth for renewals established in under 2 weeks
  • First-of-its-kind RLS solution scaled to 150+ dashboards covering 12+ user personas
  • Promoted to L5 based on scope, impact, and technical leadership

If I Were Building Today

The data infrastructure is the foundation. The next layer is intelligence on top of it: predictive insights, proactive anomaly detection, and personalized views that surface the right information to the right person before they ask. I'm already building toward this with agentic initiatives at AWS, and it's the same direction I'm taking Akaani.