Senior Software Engineer - Exchange Integrations

RAVEN; Milano; In loco; EUR 90.000 - 120.000; Tempo pieno

Descrizione del lavoro

RAVEN is seeking an experienced engineer to own end-to-end exchange integrations and low-latency trading workflows in Milan. You will design and operate high-throughput services, ensure correctness under network retries, and optimize latency across REST, WebSocket, and FIX pathways.

Collaboration across global teams is required. Ideal candidates have deep Rust or modern C++, strong Linux networking skills, and a track record in distributed systems.

Competenze

• Experience designing and operating low-latency systems with third-party services over the network.
• Proven ability to reason about distributed systems, idempotency, and reconciliation.
• Strong knowledge of Linux networking and TLS in hot-path scenarios.
• Fluent English, working across international offices.

Mansioni

• Build and operate venue integrations with centralized exchanges via REST, WebSocket, and FIX.
• Optimize quote-to-wire latency and manage rate limits to maximize throughput and reliability.
• Maintain state machines for orders and accounts, ensuring correctness under retries and replays.
• Ensure reliability through failover, degraded-mode behavior, and proactive monitoring.

Conoscenze

Low-latency systems

Distributed systems

Linux networking

Fluent English

Strumenti

Rust

C++

Descrizione del lavoro

Every decision a strategy makes ends its life as an order on an exchange, and every unit of risk we carry is only as known as our account state is fresh. You own our exchange integrations end to end. Every quote that leaves the building and every balance that comes back passes through the layer you own. A slow cancel is adverse selection with our name on it. A stale balance is risk we are carrying without knowing it.

Responsibilities

Venue integrations.

New venues from documentation to production, whatever shape they arrive in. Centralized exchanges over REST, WebSocket, and FIX - signing on the hot path, session lifecycle, cancel-on-disconnect, and every quirk the documentation forgot. DeFi venues on the chains we run nodes for (Ethereum, Arbitrum, Polygon, Avalanche), where an order is a signed transaction and nonce management, gas strategy, revert protection, and confirmation semantics decide when a fill is actually a fill. Prediction markets, where resolution and settlement are first-class events and positions track per outcome.

The quoting path.

Quote-to-wire and quote-to-ack latency, batch amends, connection warmth. Rate limits are a budget. You build the accounting so strategies spend it where it earns, and the headroom so mass-cancel always fires.

Order and account state.

State machines that survive lost acks, duplicate fills, and out-of-order updates, with idempotency built on APIs that offer none. Real-time balances and positions everywhere we hold risk - margin and collateral, on-chain balances and allowances, prediction-market positions per outcome - with drift against the venue caught in seconds, not at end of day.

Reliability.

Session failover, degraded-mode behavior, maintenance windows, and anomalies alerting into Sauron, our monitoring stack, before a trader notices.

Who You Are

• You have built and operated low-latency systems talking to third-party services over the network. Exchange connectivity is the shortest path in, but real-time bidding, telecoms signalling, and game networking build the same instincts
• Protocol work under a deadline - text and binary over TCP and WebSocket, session lifecycle and reconnects, state that stays correct when the far side misbehaves
• Deep Rust or modern C++. The stack is Rust, and you will be productive in it fast
• Distributed-systems judgment across consistency, idempotency, and reconciliation
• Linux and networking depth, including what TLS and signing cost on a hot path
• You measure before you optimize, and your numbers survive scrutiny
• Fluent English. It is the working language across both offices
• Exchange venues, blockchain execution, and prediction markets are a plus, not a filter. We teach markets to someone who brings the engineering
• No degree or certifications required. HFT pedigree is welcome; evidence is mandatory