Senior Azure Engineer - Fix Cosmos DB Throttling & Runaway Costs on Reporting Layer

Our B2B SaaS platform runs on .NET Core microservices in Azure. Our reporting/analytics layer currently queries Cosmos DB directly, and it's become a serious problem: heavy RU consumption, 429 throttling under load, degraded dashboards for customers, and a monthly Azure bill that's climbing fast. We need a senior engineer to come in, diagnose it properly, and give us a concrete fix — fast. This is a high-stakes, well-defined engagement. We don't need a junior to experiment; we need someone who's solved exactly this before and can make the right architectural call under real constraints. What we need in one week: Audit our current Cosmos DB query patterns and pinpoint what's driving RU cost and throttling (hot partitions, cross-partition scans, unindexed queries, chatty read paths) Deliver a clear remediation plan — quick wins we can ship immediately, plus the right longer-term architecture (we're open to offloading analytics to a proper reporting layer, e.g. Databricks / medallion, if that's the correct answer) Implement the highest-impact fixes within the week and measure the before/after (RU/s, cost, latency) Leave us with a short written findings + recommendations doc our team can act on Must-haves: Deep Azure + Cosmos DB experience — you understand RU economics, partitioning, and indexing at a real level Proven track record cutting cloud cost / fixing throttling on a production system (please cite a specific example) Comfortable in a live .NET Core microservices codebase Bonus: Databricks / medallion-architecture reporting, Azure Service Bus We work AI-first. We'll provide a Claude Pro seat and expect you to use AI tooling (Claude Code, Cursor) to move fast — reading the codebase, tracing query paths, and drafting the fix efficiently. Fluency here is a real plus. To apply, please answer: - Describe a time you cut Cosmos DB cost or fixed 429 throttling in production. What was the root cause and what was the before/after? - When would you move analytics off Cosmos entirely vs. optimize in place? - Can you start within 48 hours? High-priority fix. Strong outcome here leads directly into a longer engagement on our reporting roadmap.