Senior ML Researcher - Post-Training & Inference Optimizer

Thurn Partners Ltd; Miami (FL); On-site; USD 180,000 - 260,000; Full time

Job summary

Thurn Partners Ltd in the United States is seeking a highly skilled ML researcher to join post-training research. You will adapt large pretrained models for financial markets, handle noisy data and shifting distributions, and work within tight latency budgets to produce production-ready results.

You will collaborate with engineering and trading teams to take research into production, focusing on efficient inference, GPU-accelerated workflows, and real-world market outcomes rather than benchmarks.

Qualifications

• Minimum GPA 3.5/4.0 in quantitative field.
• Industry ML experience with hands-on post-training work.
• Production-quality Python and C++ coding.
• GPU memory hierarchy and kernel diagnostics knowledge.
• Eligible to work in the United States.

Responsibilities

• Fine-tune and align large models for domain tasks.
• Build evaluation measuring real-world outcomes.
• Quantize and prune to reduce latency costs.
• Write and profile custom GPU kernels when needed.
• Deploy research into production with engineering and trading teams.

Skills

Python

C++

Post-training ML

GPU knowledge

Production-grade code

Education

Bachelor's degree in a quantitative field

GPA 3.5/4.0

PhD preferred

Master's degree preferred

Tools

CUDA

Triton

Nsight

NCCL

pybind11

DeepSpeed/Megatron

Job description

Thurn Partners Ltd in the United States is seeking a highly skilled ML researcher to join post-training research. You will adapt large pretrained models for financial markets, handle noisy data and shifting distributions, and work within tight latency budgets to produce production-ready results.

You will collaborate with engineering and trading teams to take research into production, focusing on efficient inference, GPU-accelerated workflows, and real-world market outcomes rather than benchmarks.