C++ Real-Time Financial Pipeline
Budget: ₹12,500 – ₹37,500 INR
I have a set of bite-size C++ fixes and feature tweaks that need to be knocked out first, but the core of this engagement is to turn my concept for a real-time financial data pipeline into a production-ready application.
What I already know I need:
• Solid command of modern C++ (17/20), efficient data structures, and multithreading techniques.
• Experience building low-latency systems that ingest and process streaming market data.
• Comfortable persisting large time-series datasets and keeping read/write latency predictable under load—whether that’s PostgreSQL, InfluxDB, Timescale, or a comparable solution.
Pipeline scope
My priorities are the Data collection and Data processing stages. You will design the ingestion layer, wire up real-time feeds, batch or stream-clean the data, and store it so downstream services can query it without noticeable lag. Visualisation can wait; right now I need the plumbing rock solid.
Key deliverables
1. Quick turnaround on the initial set of C++ maintenance tasks (I’ll provide a detailed ticket list).
2. A modular pipeline that:
• Connects to one or more financial data sources (FIX/REST/WebSocket—exact endpoint details supplied).
• Buffers and parses the stream, applying configurable business rules in real time.
• Writes clean, timestamp-accurate records to the chosen database with back-pressure handling.
• Exposes a lightweight API stub that proves the data can be queried in near-real time.
3. Build scripts/tests so I can spin the whole stack up with a single command.
Acceptance criteria
– End-to-end latency from feed to DB insert consistently < 200 ms on my reference hardware.
– Zero data loss over a continuous 8-hour run while ingesting 5 k msg/s.
– Code passes clang-tidy and unit tests; integration tests cover reconnect scenarios.
If this matches your wheelhouse, let’s discuss the initial tasks and agree on milestones for the full build.
What I already know I need:
• Solid command of modern C++ (17/20), efficient data structures, and multithreading techniques.
• Experience building low-latency systems that ingest and process streaming market data.
• Comfortable persisting large time-series datasets and keeping read/write latency predictable under load—whether that’s PostgreSQL, InfluxDB, Timescale, or a comparable solution.
Pipeline scope
My priorities are the Data collection and Data processing stages. You will design the ingestion layer, wire up real-time feeds, batch or stream-clean the data, and store it so downstream services can query it without noticeable lag. Visualisation can wait; right now I need the plumbing rock solid.
Key deliverables
1. Quick turnaround on the initial set of C++ maintenance tasks (I’ll provide a detailed ticket list).
2. A modular pipeline that:
• Connects to one or more financial data sources (FIX/REST/WebSocket—exact endpoint details supplied).
• Buffers and parses the stream, applying configurable business rules in real time.
• Writes clean, timestamp-accurate records to the chosen database with back-pressure handling.
• Exposes a lightweight API stub that proves the data can be queried in near-real time.
3. Build scripts/tests so I can spin the whole stack up with a single command.
Acceptance criteria
– End-to-end latency from feed to DB insert consistently < 200 ms on my reference hardware.
– Zero data loss over a continuous 8-hour run while ingesting 5 k msg/s.
– Code passes clang-tidy and unit tests; integration tests cover reconnect scenarios.
If this matches your wheelhouse, let’s discuss the initial tasks and agree on milestones for the full build.