Forward Deployed Engineer - Applied AI/ML
Posted 7 days ago
The role in short
Builds and trains Physics AI models (surrogates, neural operators) for customer engineering problems and deploys them into their production workflows.
JOIN THE REVOLUTION IN ENGINEERING INNOVATION
Luminary helps engineering companies be more competitive by getting to market faster, creating better products, and reducing development risk. We do this through our Physics AI platform — the fastest and easiest way to build and deploy models that understand and instantly predict physical reality with precision. Our customers span industries from automotive and aerospace to defense, industrial, semiconductors, and energy — ranging from hyper-growth startups to Fortune 100 enterprises. Luminary is a Series B company headquartered in San Mateo, California.
YOUR IMPACT
As an Applied AI/ML Scientist on Luminary's Applied AI/ML team, you build the Physics AI models that power customer outcomes. You work in a matrix structure inside customer value delivery teams alongside a Lead Delivery Engineer, Applications Engineers, and Data & Platform engineers. You take real customer engineering problems, design and train the right model architectures, and partner with the team to deploy those models into production engineering workflows. You operate at the boundary of cutting-edge research and applied delivery — staying connected to the frontier of physics-informed ML while making sure your work ships and gets used.
WHAT YOU'LL DO
- Own model development for Physics AI engagements: architecture selection, training pipeline design, hyperparameter tuning, evaluation, and validation against ground-truth simulation.
- Work with Applications Engineers to ensure training data is physically meaningful and adequate for the target use case.
- Partner with Data & Platform engineers to operationalize training pipelines, model registries, and inference serving.
- Collaborate with Luminary Research to apply state-of-the-art techniques — neural operators, diffusion models, geometric deep learning, latent representations — to real customer problems.
- Embrace co-engineering: work side-by-side with customer data scientists and engineers, sharing methodology and building model literacy on the customer side.
- Bring back signal from delivery into Research and Product, helping shape the next generation of Luminary's Physics AI methods and platform.
- Mentor junior team members and contribute to internal best practices for applied physics-informed ML.
WHAT YOU BRING
- 5–10 years of experience in applied machine learning, with significant exposure to scientific computing, engineering simulation, or physics-informed ML. Principal-level candidates trend toward the upper end of the range.
- Strong proficiency in Python and PyTorch (or equivalent deep learning framework). You write production-quality ML code, not just research notebooks.
- Hands-on experience training and deploying models on engineering or scientific data — surrogate models, neural operators, graph neural networks, diffusion models, or related architectures.
- Working knowledge of engineering simulation: CFD, FEA, EM, thermal, or related — enough to collaborate effectively with domain experts and understand what the model needs to learn.
- Experience with distributed training, GPU workloads, and modern ML infrastructure (experiment tracking, model registries, inference serving).
- Strong scientific mindset: rigorous experimentation, careful evaluation, honest reporting of what works and what does not.
- Customer-facing presence; comfortable explaining model architectures and limitations to engineering audiences.
- Self-starter mentality, persistent through iteration, willing to travel occasionally to customer sites.
PREFERRED QUALIFICATIONS
- Advanced degree (MS or PhD) in Computer Science, Applied Math, Physics, Engineering, or related quantitative discipline.
- Published work in physics-informed ML, neural operators, scientific machine learning, or related fields.
- Experience with physics-informed AI/ML frameworks (e.g. PhysicsNeMo, JAX-based scientific ML stacks) or foundation model fine-tuning pipelines for scientific data.
- Prior experience in applied research roles at engineering, simulation, or scientific computing companies.
- Track record of shipping models into production engineering workflows.
Prep for this interview
3 questions likely to come up at Luminary Cloud, each matched to a line in this posting.
- 1Travel
They wrote: “Self-starter mentality, persistent through iteration, willing to travel occasionally to customer sites.”
How do you handle being on the road with customers for weeks at a time?
How to answer
Why they ask: Many FDE roles involve regular travel. They check you know what you're signing up for.
- Your past travel or on-site experience
- How you stay productive away from the team
- What travel level works for you, said plainly
Tip: Be honest about what works for you. A mismatch here hurts both sides later.
- 2LLMs, RAG and agents
They wrote: “Experience with physics-informed AI/ML frameworks (e.g. PhysicsNeMo, JAX-based scientific ML stacks) or foundation model fine-tuning pipelines for scientific data.”
When would you fine-tune a model, and when is prompting or retrieval enough?
How to answer
Why they ask: Customers often ask for fine-tuning first. They want to hear you pick the cheapest thing that works.
- Try clear prompts and examples first; measure against a test set
- If the model lacks the customer's facts, add retrieval
- Fine-tune when you need a format or behaviour prompts can't hold, and you have good examples
- Count the cost: training data, upkeep and re-testing on every model update
Tip: Start from what's wrong: missing knowledge points to retrieval, wrong style or format points to fine-tuning.
- 3Coding
They wrote: “Strong proficiency in Python and PyTorch (or equivalent deep learning framework). You write production-quality ML code, not just research notebooks.”
Here's a CSV export from a customer. Write a script that cleans it and loads it into our API.
How to answer
Why they ask: Very close to real FDE work. They watch how you handle bad rows and failures.
- Read with a proper CSV parser; check the header first
- Validate each row and collect errors, not crash on the first one
- Send in batches with retries and a clear report at the end
- Write one or two quick tests for the parsing
Tip: Talk through edge cases out loud: encodings, missing fields, duplicates, rate limits.
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.



