Bluedart Tracker Engine

I need an extremely high-throughput solution that can sweep through a contiguous range of Bluedart tracking numbers—up to billions at a time—and extract only the records that meet my filters. Before each run I will supply:

• the starting and ending tracking numbers
• the specific “from” city, “to” city, and/or any keywords that must appear in the shipment details

Every matching record has to be written to a single JSON file (the chosen output format), with the full tracking payload preserved so I can parse it later.

Timing is critical: the crawler must clear a range of roughly one billion numbers in no more than seven days, so please architect with aggressive parallelism, pagination handling, request throttling, and fault-tolerant retries in mind. I am open to clustered, serverless, or GPU-accelerated approaches as long as you stay within Bluedart’s public tracking endpoint restrictions and avoid IP blocks.

Deliverables
1. A runnable script or compiled service, plus clear setup instructions
2. The first test run’s JSON output proving it can filter correctly on a smaller sample
3. Documentation of the concurrency model and any rate-limit safeguards so I can scale confidently to the full data set

I will consider the job complete when the engine processes an entire billion-number range, captures every qualifying record, and outputs a valid JSON file without timeouts or data loss.