Software Engineer, Applied Machine Learning

About this role:

We are seeking a hands-on, production-focused Applied Machine Learning Engineer to take technical ownership of the model layer powering our next-generation generative media platform. In this role, you will bridge the gap between cutting-edge generative research and scalable, consumer-facing products.

You will split your time between extending SOTA open-source models with additional capabilities and helping maintain our fleet of generative model APIs.
What you’ll do:

• Novel Model Pipelines: Work with our post-training team to extend SOTA image, video, audio, and 3D models with additional capabilities and modalities. Develop novel approaches to model conditioning, generation, and editing, including both training-free methods and model fine-tuning.

• Fine-Tuning: Leverage our massive GPU fleet to fine-tune generative models for novel capabilities. Build and maintain fine-tuning APIs that allow customers to customize models for their specific needs.

• Architecture & Abstraction: Identify common patterns across the models we serve and develop reusable components, abstractions, and building blocks that accelerate the development of new model capabilities and inference pipelines.

• Inference Optimization: Work hand-in-hand with our ML Performance & Optimization team to apply state-of-the-art inference techniques and best practices, ensuring models run efficiently with low latency, high throughput, and optimal GPU utilization.

• Production Deployment: Build, deploy, and maintain scalable, reliable generative model APIs. Anticipate and resolve production challenges to ensure our models serve customers reliably at scale.

• Customer Collaboration: Work directly with customers, including some of the world's largest e-commerce retailers and film and TV production studios, to develop novel solutions to their generative media needs.

Qualifications/Nice-to-haves:

• Experience: 3+ years of professional experience as an Applied ML Engineer, with at least 1–2 years focused on generative media or computer vision.

• Core Frameworks: Expert-level proficiency in Python and PyTorch.

• Generative Media: Deep practical understanding of diffusion and flow-based generative models.

• Open-Source Tooling: Hands-on experience working with open-weight model ecosystems, including Hugging Face, Diffusers, and related tooling.

• Engineering Rigor: Ability to anticipate and solve challenges that arise when deploying ML models to production. Strong engineering judgment in designing systems that are scalable, reliable, secure, safe, and performant.

• Training-Free Model Extensions: Experience designing and implementing training-free extensions to image, video, audio, or 3D generative models, such as novel conditioning methods, inference-time modifications, or new model capabilities.

• Model Post-Training: Experience developing and executing custom post-training or fine-tuning approaches to extend generative models with additional capabilities.

• Startup Experience: Track record of working in fast-paced startup environments or digital media and entertainment industries.

• Ability to Ship: Demonstrated ability to independently take ambitious ML ideas from concept to production.

What we offer at fal:

• Interesting and challenging work

• A lot of learning and growth opportunities

• Health, dental, and vision insurance (US)

• Regular team events and offsites