AI Researcher (Defense Tech)

Remote (site in Austin,TX)

Competitive Salary + Equity

Full time

Confidential search for a deep tech company building next generation microwave and laser technologies for commercial and government applications.

This is a hands on, high ownership role for someone who can take an undefined technical problem, work out how AI/ML applies to it, and turn it into a working prototype in months rather than years.

You'll work directly with SMEs across lasers, electromagnetics, photonics and optomechanics, using agentic AI workflows and physics based simulation to explore, test and prove out new concepts before anything gets built for real.

What you'll be doing

• Take loosely defined technical problems from a technical leader, scope them, and work out how AI/ML can be applied
• Build agentic research workflows that search and synthesize literature, patents, simulation results and experimental data
• Combine AI/ML with physics based simulation to explore and optimize large numbers of design variants across areas like optical design, materials, thermal performance, beam quality and metasurfaces
• Build real prototypes, software and hardware, not just train models and hand them off
• Work horizontally across programs including HPM, directed energy/optics and pixel technology

What we're looking for

• AI/ML practitioner with a strong hard science foundation (Physics, EE, Applied Physics or similar)
• PhD attractive, but strong Master's candidates absolutely in the mix
• Applied AI background rather than pure theoretical ML research
• Track record applying AI/ML to real world scientific, engineering or national security problems
• Experience in any of: EW, SIGINT/ELINT, RF, directed energy, HEL, optical/laser physics, radar, SAR, target recognition, satellite imagery or electromagnetic systems
• Comfortable moving between software, data, physics and hardware
• A builder who wants to stay hands on rather than move into management
• Comfortable operating in uncertainty and solving problems independently