Forward Deployed Engineer, Production
Posted 11 days ago
The role in short
Embed with AI labs to run rapid experiments, build evaluation tooling, and optimize data production pipelines to directly improve customer voice models.
About David AI
At David AI, our vision is to bring impactful AI into billions of lives through voice. Our mission is to help the world's leading AI companies build better voice models and agents. We're building the data and evaluation infrastructure behind those models.
Founded in 2024, we're a fast-growing Series B company backed by NVIDIA, Meritech, First Round, YC, and others, with ~70 people across San Francisco and New York City.
About our Engineering team
At David AI, our engineers build the pipelines, platforms, and models that transform raw audio into high-signal data for leading AI labs and enterprises. We're a tight-knit team of product engineers, infrastructure specialists, and machine learning experts focused on building the world's first audio data research company.
We move fast, own our work end-to-end, and ship to production daily. Our team designs real-time pipelines handling terabytes of speech data and deploys cutting-edge generative audio models.
About this role
We generate audio data for the world’s leading speech and audio labs. Data only matters if it makes a customer’s model measurably better, and finding out whether it does is your job. As a Forward Deployed Engineer, Production, you’ll own the loop between a customer’s model, our research team, and our data production. In a typical week you’ll run an ablation on Monday, see a signal on Tuesday, change how we collect data on Wednesday, and ship a recommendation that makes it to the customer by Friday. You’ll be shipping something on the order of days, not weeks.
This role is for someone who is equally comfortable in a codebase, translating customer requirements into how we produce data, and in the details of how we ship product features to yield data that improves our customers’ models. You’ll care about why a model regressed, and also about which recording setup, product experience, or automated QA step caused it.
In this role, you will
- Run the daily experiment loop. Turn customer questions into fast, rigorous experiments, get results in days, and iterate until the answer is clear enough to act on.
- Design ablations and benchmarks that change decisions. Work with our research engineers to isolate what drives model behavior. Then translate the results into concrete changes in what we collect, how we collect it, and how we filter and weigh it.
- Develop operational intuition for data quality. Get close to how our data is produced, including recording conditions, contributor pools, annotation, and QA, so you can trace model outcomes back to operational causes and fix them at the source.
- Build the tooling that makes speed possible. Ship the eval harnesses, pipelines, and analysis tools that make the next experiment faster than the last, across every customer.
- Close the loop with the model improvement lifecycle. Make sure what you learn changes what we produce, so our data keeps improving the models that depend on it.
- Shape how we work. Help define our experimentation practices, engineering culture, and roadmap as the team grows.
Your background looks like
- AI-native coding experience with a track record of fast, impact-driven experiments and product features across the stack.
- Strong problem-solving skills and the ability to work independently or as part of a collaborative team.
- Ability to translate open-ended research questions into concrete experiments and clear recommendations.
- Excellent communication skills, with the ability to explain technical findings to both researchers and non-technical stakeholders.
- Desire to thrive in a fast-paced, dynamic environment and adapt quickly as the audio AI field evolves.
- Excited to join our team in-person in San Francisco.
Bonus points if you have
- At least 3 years of relevant engineering or product experience (product engineering, ML engineering, or applied research)
- Degree in computer science, electrical engineering, or a related technical field, or equivalent practical experience.
- Experience working with audio or speech models, or with audio/signal data more broadly.
- Familiarity with machine learning research workflows — training, fine-tuning, evaluation, and metrics.
- Experience working directly with technical or enterprise customers.
Benefits
- Unlimited PTO.
- Top-notch health, dental, and vision coverage with 100% coverage for most plans.
- FSA & HSA access.
- 401k access.
- Meals 2x daily through DoorDash + snacks and beverages available at the office.
- Unlimited company-sponsored Barry's classes.
Visas
We sponsor international candidates across several visa types, including H-1B transfers, H-1B1, O-1, E-3, and TN. We also hire candidates on STEM OPT. If you’re on STEM OPT, we’ll want to confirm you have a path to one of the visas listed above before your OPT ends. While we can’t guarantee any individual outcome, we’ve successfully sponsored teammates from all over the world.
Prep for this interview
2 questions likely to come up at David AI, each matched to a line in this posting.
- 1Discovery and scoping
They wrote: “Excellent communication skills, with the ability to explain technical findings to both researchers and non-technical stakeholders.”
A customer says “we want AI in our claims process.” What do you ask in the first meeting?
How to answer
Why they ask: Most FDE work starts this vague. They want to hear you find the real problem before you pick a tool.
- Map the current process: who touches a claim, in what order, with which systems
- Find the slowest or most error-prone step and ask what it costs them
- Agree on one measurable goal (for example time per claim, or error rate)
- Ask what data you can access, and who signs off on security
Tip: Ask about the work, not the technology. Who does it today, how long it takes, what a mistake costs, and how they'd know it got better.
- 2LLMs, RAG and agents
They wrote: “Build the tooling that makes speed possible. Ship the eval harnesses, pipelines, and analysis tools that make the next experiment faster than the last, across every customer.”
How do you know an LLM feature is good enough to ship?
How to answer
Why they ask: Evals are what separate a demo from production. Many roles list them by name.
- Write 50–200 real examples with expected outcomes, with customer experts
- Agree on a pass bar and the errors that are never acceptable
- Score automatically where possible; review a sample by hand
- Re-run on every prompt or model change, and watch live traffic after launch
Tip: Describe a test set built with the customer, a pass bar agreed up front, and checks that run on every change.
Not the one? Get the next one by email
New roles land here every day. Pick what you want to hear about.
Want the job, not just the posting? Learn how FDEs work.
FDE Info is our free weekly newsletter for forward deployed engineers: tools, field notes from real deployments, and how FDE teams hire and pay.
Free. One email a week. Unsubscribe in one click.