Senior Member of Technical Staff

Acceler8 Talent; Cambridge (MA); Ibrido; USD 120.000 - 200.000; Tempo pieno

Descrizione del lavoro

Acceler8 Talent is partnering with an early-stage team to build an AI-native physics laboratory focused on AI data-center engineering. You will design and validate physics-based models of real systems, iterating with ML surrogates to accelerate design exploration and operational decisions.

The role combines deep physical-domain expertise with practical ML, delivering reusable capabilities across engagements and contributing from problem definition to deployment.

Competenze

• Depth in fluid dynamics, heat transfer, or cooling systems.
• Strong understanding of thermodynamics and energy conversion.
• Experience with multiphysics modelling and digital twins.

Mansioni

• Build physics-based models of real engineering systems, owning assumptions and numerical methods.
• Combine first-principles simulations with learned surrogates to accelerate design explorations.
• Validate models against measurements and assess where learned models can be trusted.
• Deliver usable tools from engagements to enable repeated use by the team.
• Own work from problem definition through deployment.

Descrizione del lavoro

Boston, MA | Hybrid - Can be fully remote across the US

Build AI that helps design real physical systems.

I am partnering with an early-stage team building an AI-native physics laboratory, bringing together scientists and engineers from organizations including Google, NVIDIA, Meta, MIT and Harvard.

The initial focus is AI data-center engineering: power, cooling and the interactions between them. The longer-term ambition is to turn advances in AI into new scientific discoveries and practical engineering capabilities.

This role combines deep physical-domain expertise with practical machine learning. You’ll build models, validate them against measurements and put them into workflows engineers can use to make better design and operating decisions.

What you’ll do

• Build physics-based models of real engineering systems, owning the assumptions, boundary conditions and numerical methods.
• Combine first-principles simulations with learned surrogates to accelerate design exploration and analysis.
• Validate models against measured data, investigate disagreements and establish where learned models can be trusted.
• Apply models to customer problems, delivering usable, repeatable tools and turning each engagement into capabilities the team can reuse.
• Own your work from problem definition through implementation, validation and deployment.

What we’re looking for

You’ve personally built and validated simulations, digital twins, reduced-order models or first-principles analyses of real physical systems. You can explain which modelling choices drive the answer, where your model breaks down and what you did when results disagreed with measurements.

You’ll bring depth in a relevant domain, such as:

• Fluid dynamics, heat transfer or cooling systems.
• Thermodynamics, energy conversion or HVAC.
• Electrical power systems, power electronics or electromagnetics.
• Multiphysics modelling and industrial digital twins.

You also understand how to select and apply ML approaches to physical problems, explain why an approach fits, and assess its results. Strong domain expertise can compensate for a shorter ML track record.

A PhD or equivalent research experience is valuable; a master’s with relevant industry experience is also welcome.

Useful additional experience

PyTorch or JAX, neural operators, physics-informed networks, uncertainty quantification, design optimisation, or deploying scientific models behind APIs with automated correctness tests.

Please Note: You need to be a US Citizen or Green Card holder - no visa sponsorship available