Machine Learning Engineering Internship: Summer 2027

Susquehanna International Group, LLP; Bala Cynwyd (PA); On-site; USD 992,000 - 1,543,000; Full time

Job summary

Susquehanna International Group, LLP invites standout students to join a 10-week Machine Learning Engineering Internship. You'll build scalable training pipelines for GPU clusters, work on model inference in production trading systems, and contribute to data infrastructure that handles large, noisy market datasets.

Mentorship from senior engineers and researchers will guide you. The program emphasizes hands-on ML work, training at scale, and exposure to Susquehanna's quant and trading practices.

Qualifications

• Pursuing a Bachelor's, Master's, or PhD in a technical field with intention to graduate and begin full-time by Aug 2028.
• Strong Python programming skills; comfortable with C++ as a systems language.
• Hands-on experience with ML frameworks like PyTorch or JAX beyond API usage.

Responsibilities

• Build and optimize training pipelines across GPU infrastructure.
• Work on inference and deployment, focusing on latency and throughput in production systems.
• Develop data infrastructure to move large market datasets through ingestion and transformation.
• Profile and benchmark ML workloads; contribute to internal libraries and tools.

Skills

Python

C++

PyTorch

JAX

CUDA

Triton

DSA

Concurrency

Data processing

Open-source

Education

Bachelor's/Master's/PhD in CS/ML/EE/Math/Physics/Stats

Tools

CUDA

Triton

CuTe DSL

Job description

Overview

Our Machine Learning Engineering Internship is a 10-week immersive experience designed for students who are passionate about building the systems at the intersection of machine learning, large-scale data, and markets.

As a Machine Learning Engineering Intern, you'll work on high-impact projects that closely reflect the challenges and workflows of our full-time engineering team. You'll apply your software engineering skills to real machine learning systems while developing a deep understanding of how machine learning integrates into Susquehanna's research and trading systems. A model is useful only if it can be trained quickly, fed reliably, and run fast enough to act on - our engineers own the systems that make that true, and we scopeintern projects the same way.

Susquehanna's proprietary datasets and computing infrastructure - including a rapidly growing cluster of thousands of high-end GPUs - support computationally intensive training, simulation, and rapid experimentation. Engineering teams here are small and highly collaborative, ideas are debated openly, and work that proves out reaches production quickly.

What You Can Expect

• Build and optimize training pipelines that run across our GPU infrastructure, including distributed training for large models
• Work on inference and deployment — the latency, throughput, and cost of models running in production trading systems
• Develop the data infrastructure behind our research, moving large and noisy market datasets through ingestion, storage, and transformation
• Profile and benchmark machine learning workloads, and contribute to the internal libraries and open-source tools our researchers rely on every day
• One-on-one mentorship from experienced engineers and researchers
• Participate in a comprehensive education program with deep dives into Susquehanna's ML, quant, and trading practices
• No prior finance background required

What we’re looking for

• Currently pursuing a Bachelor's, Master's, or PhD in computer science, machine learning, electrical engineering, mathematics, physics, statistics, or a related technical field. Intention to graduate and begin full time employment by August 2028
• Strong programming skills in Python, working comfort in a systems language such as C++, a plus
• Hands‑on experience with machine learning frameworks such as PyTorch or JAX that goes beyond calling the API - training at scale, extending a library, or making an existing workflow measurably faster
• Exposure to the systems machine learning runs on: distributed training, GPU programming, orchestration, or large-scale data processing. We're more interested in what you did with a tool and what it changed than in seeing it named on a list
• Experience with GPU kernel programming in CUDA, Triton, or CuTe DSL
• Solid computer science fundamentals: data structures, algorithms, concurrency, and an understanding of how software behaves on real hardware
• A project, system, or open-source contribution you can walk through in detail - what you designed, what you decided, and what you specifically did
• Deep interest in solving complex problems and a drive to innovate in a fast-paced, competitive environment