Black Forest Labs Inc.
Forward Deployed Machine Learning Engineer
Black Forest Labs Inc., San Francisco, California, United States
Forward Deployed Machine Learning Engineer
What if the hardest part of generative AI isn't training the model, but making it work in production under constraints no one anticipated?
Our founding team pioneered Latent Diffusion and Stable Diffusion - breakthroughs that made generative AI accessible to millions. Today, our FLUX models power creative tools, design workflows, and products across industries worldwide.
Our FLUX models are best-in-class not only for their capability, but for ease of use in developing production applications. We top public benchmarks and compete at the frontier - and in most instances we're winning.
If you're relentlessly curious and driven by high agency, we want to talk.
With a team of ~50, we move fast and punch above our weight. From our labs in Freiburg - a university town in the Black Forest - and San Francisco, we're building what comes next.
What You'll Pioneer You’ll live at the intersection of cutting‑edge research and brutal production reality. Your customers won’t just want FLUX to work—they’ll need it optimised for their specific hardware, fine‑tuned for their unique use cases, and integrated into systems that weren’t designed for diffusion models in the first place.
You’ll be the person who:
Ensures FLUX models perform optimally in customer environments—whether that’s on‑premise GPU clusters or BFL‑hosted infrastructure—balancing the eternal tension between latency and output quality
Architects deep product integrations that go far beyond “here’s an API endpoint”—helping customers with everything from model hosting and deployment to inference optimisation techniques that haven’t made it into textbooks yet
Customises our foundation models for visual media to solve problems customers couldn’t articulate until you helped them understand what’s possible
Sits in technical deep‑dives with customers to diagnose performance bottlenecks, then translates those findings into solutions (and sometimes into research questions for our core team)
Discovers where generative visual AI should go next by understanding what industries are struggling with problems we could solve
Questions We’re Wrestling With
What does “optimal performance” actually mean when one customer needs 100 ms latency and another needs photorealistic quality at any cost?
How do you fine‑tune a foundation model for a customer’s specific use case without losing what made it powerful in the first place?
When should a customer run FLUX on their own infrastructure versus use our hosted solution—and how do we help them make that decision honestly?
What inference optimisations work in theory but break in production, and vice versa?
Which industries don’t yet realise they have a generative visual AI problem we could solve?
We’re figuring these out together, at the edge of what’s technically possible.
Who Thrives Here You understand diffusion models not just conceptually, but viscerally—you’ve debugged them, optimised them, served them at scale. You’ve been in the room when a customer’s integration goes wrong and you need to diagnose whether it’s a model issue, an infrastructure issue, or a fundamental misunderstanding of what the model can do.
You likely have:
Direct experience working with customers on generative AI deployment—the kind where you’re iterating on solutions in real‑time, not just following a playbook
Hands‑on expertise with generative modelling approaches, particularly finetuning, optimising and serving deep learning models in production environments
A proven track record as an ML engineer who’s shipped models that real systems depend on
Strong Python skills and intuitive understanding of API design (because demos and prototypes are how you communicate what’s possible)
The ability to explain why a diffusion model is slow to an executive and how to fix it to an engineer—in the same meeting
We’d be especially excited if you:
Have deep knowledge of diffusion models and/or flow matching, including finetuning and distillation techniques that go beyond standard tutorials
Know the FLUX ecosystem intimately—ComfyUI, common training frameworks, the tools practitioners actually use
Have battle‑tested experience optimising inference for transformer‑based models (and the scars to prove it)
Can architect solutions in complex enterprise environments where “just add more GPUs” isn’t an option
Contribute to open‑source projects in the diffusion model space and understand the community
Have deployed models on cloud platforms using state‑of‑the‑art serving infrastructure
What We’re Building Toward We’re not just supporting customers—we’re learning what it actually takes to bring frontier generative AI into production at scale. Every customer integration teaches us something we didn’t know. Every optimisation challenge reveals gaps in our understanding. If that sounds more compelling than having all the answers documented, we should talk.
Base Annual Salary for SF based role:
$180,000–$300,000 USD (depending on experience)
We’re based in Europe and value depth over noise, collaboration over hero culture, and honest technical conversations over hype. Our models have been downloaded hundreds of millions of times, but we’re still a ~50‑person team learning what’s possible at the edge of generative AI.
#J-18808-Ljbffr
Our founding team pioneered Latent Diffusion and Stable Diffusion - breakthroughs that made generative AI accessible to millions. Today, our FLUX models power creative tools, design workflows, and products across industries worldwide.
Our FLUX models are best-in-class not only for their capability, but for ease of use in developing production applications. We top public benchmarks and compete at the frontier - and in most instances we're winning.
If you're relentlessly curious and driven by high agency, we want to talk.
With a team of ~50, we move fast and punch above our weight. From our labs in Freiburg - a university town in the Black Forest - and San Francisco, we're building what comes next.
What You'll Pioneer You’ll live at the intersection of cutting‑edge research and brutal production reality. Your customers won’t just want FLUX to work—they’ll need it optimised for their specific hardware, fine‑tuned for their unique use cases, and integrated into systems that weren’t designed for diffusion models in the first place.
You’ll be the person who:
Ensures FLUX models perform optimally in customer environments—whether that’s on‑premise GPU clusters or BFL‑hosted infrastructure—balancing the eternal tension between latency and output quality
Architects deep product integrations that go far beyond “here’s an API endpoint”—helping customers with everything from model hosting and deployment to inference optimisation techniques that haven’t made it into textbooks yet
Customises our foundation models for visual media to solve problems customers couldn’t articulate until you helped them understand what’s possible
Sits in technical deep‑dives with customers to diagnose performance bottlenecks, then translates those findings into solutions (and sometimes into research questions for our core team)
Discovers where generative visual AI should go next by understanding what industries are struggling with problems we could solve
Questions We’re Wrestling With
What does “optimal performance” actually mean when one customer needs 100 ms latency and another needs photorealistic quality at any cost?
How do you fine‑tune a foundation model for a customer’s specific use case without losing what made it powerful in the first place?
When should a customer run FLUX on their own infrastructure versus use our hosted solution—and how do we help them make that decision honestly?
What inference optimisations work in theory but break in production, and vice versa?
Which industries don’t yet realise they have a generative visual AI problem we could solve?
We’re figuring these out together, at the edge of what’s technically possible.
Who Thrives Here You understand diffusion models not just conceptually, but viscerally—you’ve debugged them, optimised them, served them at scale. You’ve been in the room when a customer’s integration goes wrong and you need to diagnose whether it’s a model issue, an infrastructure issue, or a fundamental misunderstanding of what the model can do.
You likely have:
Direct experience working with customers on generative AI deployment—the kind where you’re iterating on solutions in real‑time, not just following a playbook
Hands‑on expertise with generative modelling approaches, particularly finetuning, optimising and serving deep learning models in production environments
A proven track record as an ML engineer who’s shipped models that real systems depend on
Strong Python skills and intuitive understanding of API design (because demos and prototypes are how you communicate what’s possible)
The ability to explain why a diffusion model is slow to an executive and how to fix it to an engineer—in the same meeting
We’d be especially excited if you:
Have deep knowledge of diffusion models and/or flow matching, including finetuning and distillation techniques that go beyond standard tutorials
Know the FLUX ecosystem intimately—ComfyUI, common training frameworks, the tools practitioners actually use
Have battle‑tested experience optimising inference for transformer‑based models (and the scars to prove it)
Can architect solutions in complex enterprise environments where “just add more GPUs” isn’t an option
Contribute to open‑source projects in the diffusion model space and understand the community
Have deployed models on cloud platforms using state‑of‑the‑art serving infrastructure
What We’re Building Toward We’re not just supporting customers—we’re learning what it actually takes to bring frontier generative AI into production at scale. Every customer integration teaches us something we didn’t know. Every optimisation challenge reveals gaps in our understanding. If that sounds more compelling than having all the answers documented, we should talk.
Base Annual Salary for SF based role:
$180,000–$300,000 USD (depending on experience)
We’re based in Europe and value depth over noise, collaboration over hero culture, and honest technical conversations over hype. Our models have been downloaded hundreds of millions of times, but we’re still a ~50‑person team learning what’s possible at the edge of generative AI.
#J-18808-Ljbffr