Machine Learning Engineer – Deep Learning for HFT

Empower Partners Singapore

Singapore

On-site

SGD 180,000 - 260,000

Full time

14 days+

Get a reply from this employer

Your resume and cover letter, tailored to exactly what they’re hiring for.

Get past ATS filters

Job summary

Empower Partners Singapore is seeking an AI / Machine Learning Engineer to develop and deploy deep learning models for high-frequency trading and market making. You will work on neural networks trained on raw tick-level order-book data across markets, with ownership of data prep, model research, and training pipelines.

Reporting to the Head of HFT, you will integrate models into a latency-constrained C++ platform, optimize inference, and collaborate with researchers and engineers on scalable GPU

Qualifications

  • Master’s or PhD in a quantitative field (CS/ML/Math/Physics/EE) or equivalent exceptional ability.
  • Typically 3–10 years post-degree experience in ML/AI research or engineering.
  • Recent PhD graduates with substantial hands-on model training encouraged to apply.
  • Experience owning training of deep learning models at scale (multi-GPU / >100 GB data).
  • Strong Python skills with PyTorch or JAX; proficiency in C++, Rust or CUDA.
  • Experience deploying models under latency or throughput constraints.
  • Rigorous experimental track record (first-author papers or credible benchmarks).
  • Strong mathematical foundations; ability to design experiments and assess generalisation.
  • Experience in sequence modelling and time-series prediction is relevant.
  • Knowledge of CUDA / Triton kernels and distributed training is advantageous.
  • Interest in applying deep learning to financial markets; prior trading experience welcome but not required.

Responsibilities

  • Research and train deep learning models on large-scale order-book and market data to predict short-horizon market behaviour.
  • Develop sequence and time-series models, exploring architectures and training methods suited to noisy, rapidly changing financial data.
  • Own data preparation, feature and label construction, training pipelines and reproducible experiments.
  • Evaluate models against robust baselines using time-aware out-of-sample validation, with careful controls for data leakage and overfitting.
  • Improve training efficiency and scalability through GPU computing, distributed training and performance optimisation.
  • Work closely with HFT researchers and engineers to optimise inference and integrate models into a latency-constrained C++ market-making system.
  • Evaluate production model behaviour and use live feedback to guide further research.

Skills

Python
PyTorch
JAX
C++
Rust
CUDA

Education

Master's degree
PhD

Tools

CUDA
Triton

Job description

About Our Client

Our client is an established quantitative trading firm seeking an AI / Machine Learning Engineer to develop and deploy deep learning models for high-frequency trading and market making. Reporting to the Head of HFT, you will work on large neural networks trained on raw tick-level order-book data across multiple markets and venues. The role offers ownership across data preparation, model research, training infrastructure and integration into a live C++ trading platform. You will join a small team with minimal hierarchy, direct access to decision-makers and meaningful ownership of your work. The firm offers top-of-market compensation with P&L-linked upside. We welcome experienced researchers and engineers from quantitative finance, technology companies and AI research labs, as well as exceptional recent PhD graduates with strong practical experience. Prior financial markets experience is not essential.

Job Responsibilities
  • Research and train deep learning models on large-scale order-book and market data to predict short-horizon market behaviour.
  • Develop sequence and time-series models, exploring architectures and training methods suited to noisy, rapidly changing financial data.
  • Own data preparation, feature and label construction, training pipelines and reproducible experiments.
  • Evaluate models against robust baselines using time-aware out-of-sample validation, with careful controls for data leakage and overfitting.
  • Improve training efficiency and scalability through GPU computing, distributed training and performance optimisation.
  • Work closely with HFT researchers and engineers to optimise inference and integrate models into a latency-constrained C++ market-making system.
  • Evaluate production model behaviour and use live feedback to guide further research.
Work Experience and Requirements
  • A Master’s or PhD in Computer Science, Machine Learning, Mathematics, Physics, Electrical Engineering or a related quantitative discipline, or equivalent evidence of exceptional research and engineering ability.
  • Typically 3–10 years of relevant post-degree experience.
  • Exceptional recent PhD graduates with substantial hands-on model training and implementation experience are encouraged to apply.
  • Personal ownership of training deep learning models at scale, demonstrated through datasets exceeding 100 GB or multi-GPU training.
  • Strong Python skills and regular hands-on use of PyTorch or JAX, together with proficiency in C++, Rust or CUDA.
  • Strong hands-on experience implementing and optimising deep learning models.
  • Experience deploying models under latency or throughput constraints is preferred; recent PhD graduates may demonstrate relevant ability through research implementations, benchmarking or industry internships.
  • A rigorous experimental research track record, demonstrated through first-author publications at leading ML conferences, Kaggle Master status or above, or industry research evaluated against credible baselines.
  • Strong mathematical and statistical foundations, with the ability to design experiments, interpret results and assess model generalisation.
  • Experience in sequence modelling, time-series prediction or efficient deep learning is particularly relevant.
  • Custom CUDA / Triton kernels, distributed training, inference optimisation, or order-book and market microstructure experience would be advantageous.
  • An interest in applying deep learning to financial markets.
  • Prior trading experience is welcome but not required.
Get your free, confidential resume review.

or drag and drop your file here.

Similar jobs

Similar jobs worth comparing

Machine Learning Quantitative Researcher – Crypto Derivatives
Machine Learning Quantitative Researcher – Crypto Derivatives

Unchain Data • Singapore

On-site
SGD 180,000 - 250,000
Deep Learning Engineer – HFT, Low-Latency Market Making
Deep Learning Engineer – HFT, Low-Latency Market Making

Empower Partners Singapore • Singapore

On-site
SGD 180,000 - 260,000
Machine Learning Quantitative Researcher – Crypto Derivatives
Machine Learning Quantitative Researcher – Crypto Derivatives

Hytech • Singapore

On-site
SGD 180,000 - 290,000
Machine Learning Quantitative Researcher - Crypto Derivatives
Machine Learning Quantitative Researcher - Crypto Derivatives

Fuku • Singapore

On-site
SGD 180,000 - 240,000
Machine Learning Quantitative Researcher - Crypto Derivatives
Machine Learning Quantitative Researcher - Crypto Derivatives

Fuku • Singapore

On-site
SGD 180,000 - 280,000
Quantitative Trading & Research, Quantitative Developer - Associate
Quantitative Trading & Research, Quantitative Developer - Associate

JPMorgan Chase & Co. • Singapore

On-site
SGD 120,000 - 240,000
ML Quant Researcher: Crypto Market-Making & Trading
ML Quant Researcher: Crypto Market-Making & Trading

Fuku by Momofuku • Singapore

On-site
SGD 180,000 - 280,000
Lead AI Engineer
Lead AI Engineer

Ambition • Singapore

On-site
SGD 180,000 - 280,000
Deep Learning Quantitative Researcher
Deep Learning Quantitative Researcher

Millennium • Singapore

On-site
SGD 180,000 - 260,000
Quantitative Trading & Research - AI Scientist, Quantitative Research - Associate
Quantitative Trading & Research - AI Scientist, Quantitative Research - Associate

JPMorgan Chase & Co. • Singapore

On-site
SGD 180,000 - 260,000