Acceler8 Talent
Senior RL Research Scientist
Acceler8 Talent, San Francisco, California, United States, 94199
Senior RL Research Scientist – San Francisco / Redwood City (In‑person)
Physical Superintelligence:
AI that can ask any scientific question about the physical world and get a real answer in days, not years—by pairing large models with massive automated “experiment factories.”
About the role:
Responsibilities
Own RL for real‑world tool control and long‑horizon workflows across biology, chemistry, materials, and robotics.
Build RL environments grounded in physics, simulation, and digital twins—not toy benchmarks.
Design safe / constrained RL + verifier‑integrated rewards to keep agents reliable in high‑stakes domains.
Drive offline → online transitions and evaluation pipelines as agents learn from and act on real telemetry.
Partner across LLM, agent systems, simulation, and physical‑tooling teams to deploy policies into production workflows.
High‑ownership algorithm + systems work where you shape the foundations of physical‑AI autonomy.
Work at the frontier of real‑world impact—from multi‑cancer early detection to general scientific discovery.
Operate on a uniquely scaled platform: 20k+ sq ft automated facilities producing real experimental feedback loops.
Benefits
Competitive salary + meaningful equity.
Tiny, world‑class team with a “best‑person‑you’ve‑worked‑with” talent bar.
In‑person, high‑velocity environment with deep collaboration across SF + Redwood City.
#J-18808-Ljbffr
AI that can ask any scientific question about the physical world and get a real answer in days, not years—by pairing large models with massive automated “experiment factories.”
About the role:
Responsibilities
Own RL for real‑world tool control and long‑horizon workflows across biology, chemistry, materials, and robotics.
Build RL environments grounded in physics, simulation, and digital twins—not toy benchmarks.
Design safe / constrained RL + verifier‑integrated rewards to keep agents reliable in high‑stakes domains.
Drive offline → online transitions and evaluation pipelines as agents learn from and act on real telemetry.
Partner across LLM, agent systems, simulation, and physical‑tooling teams to deploy policies into production workflows.
High‑ownership algorithm + systems work where you shape the foundations of physical‑AI autonomy.
Work at the frontier of real‑world impact—from multi‑cancer early detection to general scientific discovery.
Operate on a uniquely scaled platform: 20k+ sq ft automated facilities producing real experimental feedback loops.
Benefits
Competitive salary + meaningful equity.
Tiny, world‑class team with a “best‑person‑you’ve‑worked‑with” talent bar.
In‑person, high‑velocity environment with deep collaboration across SF + Redwood City.
#J-18808-Ljbffr