Founding Engineer, ML Inference

Reactor; San Francisco (CA); On-site; Full time

Benefits offered by this job

Competitive salary

Early equity

Health, dental, and vision coverage

Relocation support

Job summary

A media technology company in San Francisco is seeking a Founding Engineer specializing in ML Inference. This highly technical role requires expertise in the ML infrastructure stack and aims to optimize generative media performance. The ideal candidate will drive innovations in real-time model performance, design in-house inference runtimes, and optimize models through advanced techniques. Competitive salary and relocation support are offered, along with generous health coverage.

Qualifications

• Strong foundation in systems programming with a track record of identifying and resolving bottlenecks.
• Deep expertise in ML infrastructure stack, including PyTorch, TensorRT, and more.
• Knowledge of model compilation and quantization techniques.

Responsibilities

• Drive performance for real-time model performance for diffusion models.
• Design a high-performance in-house inference runtime.
• Optimize models for inference through quantization and pruning.
• Benchmark model performance to identify bottlenecks.

Skills

Systems programming

PyTorch

TensorRT

Model compilation

Quantization

NVIDIA GPU hardware knowledge

Transformer architectures

Job description

We're building a future where anyone can create interactive media applications that delight, educate, and simulate. Building a new kind of platform for real-time generative media, enabling developers to go from idea to immersive, dynamic experience in seconds. Join a small, focused team of YC and unicorn founders and senior engineers with deep expertise in 3D, generative video, developer platforms, and creative tool, aspiring to continuously push the boundaries of what's possible.

About the Role

We're looking for a Founding Engineer, ML Inference with deep expertise in high-performance ML engineering. This is a highly technical, high-impact role focused on squeezing every drop of performance from generative media models.

You'll work across the model-serving stack, designing novel inference frameworks, optimizing inference performance, and shaping the competitive edge in ultra-low-latency, high-throughput environments.

What You'll Do

• Drive our frontier position on real-time model performance for diffusion models
• Design and implement a high-performance in-house inference runtime
• Implement optimizations using torch.compile, custom CUDA kernels, and specialized inference frameworks
• Optimize neural network models for inference through quantization, pruning, and architectural modifications while maintaining accuracy
• Profile and benchmark model performance to identify computational bottlenecks
• Collaborate directly with model partner teams to directly integrate their models into our platform

Required Skills

• Strong foundation in systems programming, with a track record of identifying and resolving bottlenecks
• Deep expertise in the ML infrastructure stack: PyTorch, TensorRT, TransformerEngine, Nsight, ONNX Runtime
• Model compilation, quantization (INT8/FP16), and advanced serving architectures
• Working knowledge of GPU hardware (NVIDIA) and the ability to dive deep into the stack as needed
• Strong understanding of transformer architectures and modern ML model optimization techniques

Logistics

We are based in-person in San Francisco. We believe the best ideas and work come from being together.

• Competitive San Francisco salary and meaningful early equity.

• We sponsor visas. We are committed to working through the process together for the right candidates. If you're currently outside the US, we're also committed to helping you relocate to the US throughout this process.

• We offer generous health, dental, and vision coverage, and relocation support as needed.