Senior/Staff ML Engineer

At Quilter, we are helping electrical engineers save time and accomplish more by automating the tedious and time-consuming task of designing printed circuit boards (PCBs). Our small team is composed of experts in electrical engineering, electromagnetic simulation, ML/AI, and high-performance computing (HPC). We are inventing and leveraging novel techniques to solve the decades-old problem of automating circuit board design where today hundreds of billions of dollars are spent. We have raised $25 million in Series B funding from some of the very best and are charging full-speed toward our goal.

No matter where we come from, we're united by a common vision for the future and a core set of values we think will get us there:

1. Focus on the mission

2. Build great things that help humans

3. Demonstrate grit

4. Never stop learning

5. Pursue excellence

We're looking for a Senior/Staff ML Systems Engineer to join Quilter's ML Team. This role spans the full ML lifecycle — from problem formulation and data pipeline design through distributed training, production deployment, and ongoing model maintenance. You should be comfortable both implementing systems and reasoning about the mathematics behind them.

As one of our early engineers, you'll have significant ownership and influence over the direction of our product, architecture, and team culture.

• Design and implement end-to-end ML pipelines: data creation and curation, training, evaluation, deployment, and continuous improvement

• Build and operate high-performance inference servers for low-latency PCB layout generation

• Build distributed training infrastructure (multi-GPU, multi-node) for large-scale geometric datasets

• Build and maintain ML CI/CD systems for model validation (accuracy, latency, I/O) and continuous delivery

• Build tooling for A/B testing, controlled rollouts, and distribution drift detection

• Build automated retraining pipelines to keep production models current

• Implement and iterate on SL, SSL, and RL algorithms for geometric and PCB layout problems

• Collaborate on model architecture decisions and data representation design

• Optimize for GPU utilization, training throughput, and inference latency

We need someone who has owned the full ML lifecycle — not contributed to pieces of it, but owned it end to end.

• Production-grade Python expertise.

• Full-lifecycle ML ownership. You have personally taken at least one ML system from problem definition (choosing the architecture, designing the data format) through production deployment and ongoing maintenance. You can speak concretely about the decisions you made and why

• Training at scale. You have trained large models on large datasets (1M+ samples) using distributed training across multiple GPUs and nodes. The framework doesn't matter — PyTorch, JAX, TensorFlow, Julia, raw CUDA — as long as you wrote the code that defined and trained the model. You've dealt with the practical problems — gradient instabilities, memory constraints, slow convergence — and can describe how you solved them

• Production ML systems. You have built and maintained production inference servers, model versioning, CI/CD for ML, monitoring, and automated retraining. You know what Good looks like because you've built it first-hand. You have deployed a model you trained and kept it reliable in an automated fashion

• Mathematical fluency. You are comfortable with the math behind the models you build — optimization, probability, linear algebra, and the specifics of whatever architectures and algorithms you've worked with. You can engage in research-level conversations about novel approaches, not just implement known patterns

• Data engineering at scale. You have built or substantially contributed to data generation, cleaning, and curation pipelines handling 1M+ samples. You understand how data quality and format decisions shape model behavior

• Strong experience with ML pipeline orchestration (Kubeflow, MLflow, or similar)

• Familiarity with hardware acceleration (CUDA, TensorRT) and memory optimization techniques (gradient checkpointing, mixed precision)

• Background in cluster management and job scheduling systems

• Familiarity with cloud platforms (AWS, GCP, or Azure) for compute, storage, and ML services

• Strong communication and collaboration skills

Be prepared to walk through a system you owned end to end and discuss the technical decisions behind it.

• Experience with geometric or spatial data (point clouds, meshes, graphs, layouts)

• Experience with reinforcement learning, particularly combinatorial or constrained optimization problems

• Kubernetes experience (production deployments, scaling, monitoring)

• Infrastructure as code (Terraform, Helm)

• Container optimization for ML workloads

• Profiling and debugging tools for ML workloads (NVIDIA Nsight, PyTorch Profiler, Weights & Biases)

• Model compression techniques (knowledge distillation, pruning, quantization)

Please note: We are an equal opportunity employer. At this time, we are focused on hiring primarily within the US, with occasional exception to accommodate exceptional talent.

• Interesting and challenging work

• Competitive salary and equity benefits

• Health, dental, and vision insurance

• Regular team events and offsites (~2x / year)

• Unlimited paid time off

• Paid parental leave

Want to learn more about Quilter, our vision, and our investors? Visit our About page and visit our Blog.