Senior ML Engineer: Hardware-Optimized Inference

Fintal Partners; Chicago (IL); On-site; USD 140,000 - 200,000; Full time

Job summary

Fintal Partners is building next-generation ML infrastructure by deploying models on custom hardware. You will co-design ML models with researchers and engineers, balancing latency, precision, and resource use as core design considerations.

You will help shape the roadmap for custom hardware platforms, translating ML workloads into hardware decisions and partnering closely with hardware engineers to move from prototypes to production deployments.

Qualifications

• Deep understanding of hardware architectures for ML workloads.
• Experience mapping ML workloads to accelerators (FPGA/ASIC/other).
• Familiarity with ML frameworks (PyTorch, TensorFlow).
• Proficient in Python and C++.
• Strong collaboration across hardware, software, and research teams.

Responsibilities

• Co-design ML models with researchers and engineers while considering latency and precision constraints.
• Shape roadmap for custom hardware platforms translating ML workloads into architecture decisions.
• Work with hardware engineers to deploy ML inference solutions from research to production.
• Evaluate research on neural architecture search, quantization, ML systems, and hardware optimization.
• Drive performance optimization balancing accuracy with latency and throughput.

Skills

Hardware design awareness

Python

C++

Communication skills

ML fundamentals

Education

Master’s or PhD in EE/CS/Physics

Tools

VHDL

SystemVerilog

HLS

hls4ml

FINN

Vitis AI

Job description

Fintal Partners is building next-generation ML infrastructure by deploying models on custom hardware. You will co-design ML models with researchers and engineers, balancing latency, precision, and resource use as core design considerations.

You will help shape the roadmap for custom hardware platforms, translating ML workloads into hardware decisions and partnering closely with hardware engineers to move from prototypes to production deployments.