Machine Learning Engineer - Altas Technologies at IMC

At IMC, technology is at the core of everything we do. Our innovative in-house software drives millions of trading decisions daily, allowing us to outperform the competition through faster, more effective decision-making.

In 2023, IMC acquired Altas Technologies, a rapidly growing algorithmic trading firm with the ambition of creating tomorrow's most technically advanced trading stack. By combining their highly advanced trading strategies with IMC's execution and scaling capabilities, this move strengthens our position in the market and is an important step towards ensuring stability and long-term growth for the firm.

The Role

You will be part of a small group of engineers supporting a larger group of researchers. That ratio is deliberate — it forces us to build high quality, low overhead systems. We do not have the headcount to firefight low quality systems. Everything we build, we own end-to-end: we design it, build it, and maintain it.

The surface area is large: real-time ML inference driving trading strategies, the data infrastructure that feeds it at scale, and the research platform that enables quant researchers to find alpha. If you want to specialise in one area and hand everything else off, this isn't the role. We go deep into every part of it. That's the challenge, that's what makes it interesting, and there is always more to learn.

How We Work

Competence is the baseline here, not the differentiator. Individual output matters, but what sets people apart is the quality of their ideas and how they work together — through collaboration, discussion, and mutual respect. No egos.

Responsibilities

• Own and evolve the real-time inference system — serving ML predictions at sub-millisecond latency as part of a production trading system.
• Build and maintain petabyte-scale data and ML infrastructure — high-throughput ingestion into our datalake, end-to-end ML pipeline orchestration, and everything in between.
• Enable alpha research and bring it to production — build the research platform researchers depend on daily: fast loading of TB-scale datasets, horizontally scalable experiment compute, feature engineering, and backtesting. Work with researchers to get their ideas into production at speed, without cutting corners.

What We Expect From You

You write a systems language — preferably Rust or C++ — and you write Python. You need both to work across the entire stack.

You've built production systems before — not prototypes, not dashboards, not things that someone else operates. Systems with uptime requirements, latency budgets, and consequences when they fail.

We're looking for T-shaped engineers: deep in at least one domain — whether that's ultra-low-latency systems, ML infrastructure, high-throughput data engineering, or something adjacent — and capable across the rest.

You've likely done this in a small team before: a trading firm, a startup, a game studio, or a small team inside a larger company. A trading background is not required — some of us didn't have one either.

• 3+ years building and operating ML production systems professionally.
• Strong in Python and a systems language (Rust, C++, or similar).
• Comfortable with asynchronous, concurrent, and multi-threaded programming in high-throughput, low-latency systems.
• Production experience with ML frameworks (JAX, PyTorch, ONNX) — not just training runs, but serving and integration.
• Experience with high-throughput data pipelines, scheduling, and feature computation at scale.
• Solid foundation in mathematics and statistics — you'll need it to work effectively with researchers and reason about the models you're putting into production.
• Naturally curious, willing to go deep, and driven to take ownership. You hold yourself accountable for the quality of what you deliver.