diff --git a/docs/source/tutorials/xrt-blop-demo.ipynb b/docs/source/tutorials/xrt-blop-demo.ipynb
index 433ff8d..1b5fcc8 100644
--- a/docs/source/tutorials/xrt-blop-demo.ipynb
+++ b/docs/source/tutorials/xrt-blop-demo.ipynb
@@ -41,13 +41,13 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import sys, os\n",
+    "import time\n",
+    "\n",
     "from matplotlib import pyplot as plt\n",
-    "from blop.sim.xrt_beamline import Beamline\n",
     "\n",
-    "from blop import DOF, Objective, Agent\n",
+    "from blop import DOF, Agent, Objective\n",
     "from blop.digestion import beam_stats_digestion\n",
-    "import time"
+    "from blop.sim.xrt_beamline import Beamline"
    ]
   },
   {
@@ -78,10 +78,10 @@
     "dofs = [\n",
     "    DOF(description=\"KBV R\",\n",
     "        device=beamline.kbv_dsv,\n",
-    "        search_domain=(R1-dR1, R1+dR1)),\n",
+    "        search_domain=(R1 - dR1, R1 + dR1)),\n",
     "    DOF(description=\"KBH R\",\n",
     "        device=beamline.kbh_dsh,\n",
-    "        search_domain=(R2-dR2, R2+dR2)),\n",
+    "        search_domain=(R2 - dR2, R2 + dR2)),\n",
     "\n",
     "]"
    ]
@@ -94,7 +94,7 @@
    "outputs": [],
    "source": [
     "objectives = [\n",
-    "    Objective(name=\"bl_det_sum\", \n",
+    "    Objective(name=\"bl_det_sum\",\n",
     "              target=\"max\",\n",
     "              transform=\"log\",\n",
     "              trust_domain=(20, 1e12)),\n",
diff --git a/src/blop/agent.py b/src/blop/agent.py
index c86efa8..bbc5965 100644
--- a/src/blop/agent.py
+++ b/src/blop/agent.py
@@ -5,7 +5,7 @@
 import warnings
 from collections import OrderedDict
 from collections.abc import Callable, Generator, Hashable, Iterator, Mapping, Sequence
-from typing import Any, cast, Optional, Union
+from typing import Any, cast
 
 import bluesky.plan_stubs as bps  # noqa F401
 import botorch  # type: ignore[import-untyped]
@@ -148,7 +148,7 @@ def random_ref_point(self) -> ArrayLike:
             raise RuntimeError("'random_ref_point' is not defined for multi-objective optimization.")
         return train_targets[self.argmax_best_f(weights="random")]
 
-    def raw_inputs(self, index: Optional[Union[str, int]] = None, **subset_kwargs) -> torch.Tensor:
+    def raw_inputs(self, index: str | int | None = None, **subset_kwargs) -> torch.Tensor:
         """
         Get the raw, untransformed inputs for a DOF (or for a subset).
         """
@@ -156,7 +156,7 @@ def raw_inputs(self, index: Optional[Union[str, int]] = None, **subset_kwargs) -
             return torch.stack([self.raw_inputs(dof.name) for dof in self.dofs(**subset_kwargs)], dim=-1)
         return torch.tensor(self._table.loc[:, self.dofs[index].name].values, dtype=torch.double)
 
-    def train_inputs(self, index: Optional[Union[str, int]] = None, **subset_kwargs) -> torch.Tensor:
+    def train_inputs(self, index: str | int | None = None, **subset_kwargs) -> torch.Tensor:
         """
         A two-dimensional tensor of all DOF values for training on.
         """
@@ -168,7 +168,7 @@ def train_inputs(self, index: Optional[Union[str, int]] = None, **subset_kwargs)
         raw_inputs = self.raw_inputs(index=index, **subset_kwargs)
         return dof._transform(raw_inputs)
 
-    def raw_targets_dict(self, index: Optional[Union[str, int]] = None, **subset_kwargs) -> dict[str, torch.Tensor]:
+    def raw_targets_dict(self, index: str | int | None = None, **subset_kwargs) -> dict[str, torch.Tensor]:
         """
         Get the raw, untransformed targets for an objective (or for a subset of objectives) as a dict.
         """
@@ -177,13 +177,13 @@ def raw_targets_dict(self, index: Optional[Union[str, int]] = None, **subset_kwa
         key = self.objectives[index].name
         return {key: torch.tensor(self._table.loc[:, key].values, dtype=torch.double)}
 
-    def raw_targets(self, index: Optional[Union[str, int]] = None, **subset_kwargs) -> torch.Tensor:
+    def raw_targets(self, index: str | int | None = None, **subset_kwargs) -> torch.Tensor:
         """
         Get the raw, untransformed targets for an objective (or for a subset of objectives) as a tensor.
         """
         return torch.stack(list(self.raw_targets_dict(index=index, **subset_kwargs).values()), axis=-1)
 
-    def train_targets_dict(self, index: Optional[Union[str, int]] = None, **subset_kwargs) -> dict[str, torch.Tensor]:
+    def train_targets_dict(self, index: str | int | None = None, **subset_kwargs) -> dict[str, torch.Tensor]:
         """
         Returns the values associated with an objective name.
         """
@@ -205,7 +205,7 @@ def train_targets_dict(self, index: Optional[Union[str, int]] = None, **subset_k
 
         return targets_dict
 
-    def train_targets(self, index: Optional[Union[str, int]] = None, **subset_kwargs) -> torch.Tensor:
+    def train_targets(self, index: str | int | None = None, **subset_kwargs) -> torch.Tensor:
         """
         Returns the values associated with an objective name as an (n_samples, n_objective) tensor.
         """
diff --git a/src/blop/dofs.py b/src/blop/dofs.py
index 1b6dd8d..5266ac3 100644
--- a/src/blop/dofs.py
+++ b/src/blop/dofs.py
@@ -2,7 +2,7 @@
 import time as ttime
 import uuid
 from collections.abc import Iterable, Sequence
-from typing import Any, Literal, Optional, Union, cast, overload
+from typing import Any, Literal, cast, overload
 
 import numpy as np
 import pandas as pd
@@ -89,16 +89,16 @@ def __init__(
         name: str = None,
         description: str = "",
         type: Literal["continuous", "binary", "ordinal", "categorical"] = "continuous",
-        search_domain: Union[tuple[float, float], set[int], set[str], set[bool]] = (-np.inf, np.inf),
-        trust_domain: Optional[Union[tuple[float, float], set[int], set[str], set[bool]]] = None,
-        domain: Optional[Union[tuple[float, float], set[int], set[str], set[bool]]] = None,
+        search_domain: tuple[float, float] | set[int] | set[str] | set[bool] = (-np.inf, np.inf),
+        trust_domain: tuple[float, float] | set[int] | set[str] | set[bool] | None = None,
+        domain: tuple[float, float] | set[int] | set[str] | set[bool] | None = None,
         active: bool = True,
         read_only: bool = False,
-        transform: Optional[Literal["log", "logit", "arctanh"]] = None,
-        device: Optional[Signal] = None,
+        transform: Literal["log", "logit", "arctanh"] | None = None,
+        device: Signal | None = None,
         tags: list[str] = None,
         travel_expense: float = 1,
-        units: Optional[str] = None,
+        units: str | None = None,
     ):
         # these should be set first, as they are just variables
         self.name = name
@@ -184,7 +184,7 @@ def __repr__(self) -> str:
         return f"{self.__class__.__name__}({filling})"
 
     @property
-    def search_domain(self) -> Union[tuple[float, float], set[int], set[str], set[bool]]:
+    def search_domain(self) -> tuple[float, float] | set[int] | set[str] | set[bool]:
         """
         A writable DOF always has a search domain, and a read-only DOF will return its current value.
         """
@@ -198,21 +198,21 @@ def search_domain(self) -> Union[tuple[float, float], set[int], set[str], set[bo
             return self._search_domain
 
     @search_domain.setter
-    def search_domain(self, value: Union[tuple[float, float], set[int], set[str], set[bool]]):
+    def search_domain(self, value: tuple[float, float] | set[int] | set[str] | set[bool]):
         """
         Make sure that the search domain is within the trust domain before setting it.
         """
         value = validate_set(value, type=self.type)
         trust_domain = self.trust_domain
         if is_subset(value, trust_domain, type=self.type, proper=False):
-            self._search_domain = cast(Union[tuple[float, float], set[int], set[str], set[bool]], value)
+            self._search_domain = cast(tuple[float, float] | set[int] | set[str] | set[bool], value)
         else:
             raise ValueError(
                 f"Cannot set search domain to {value} as it is not a subset of the trust domain {trust_domain}."
             )
 
     @property
-    def trust_domain(self) -> Union[tuple[float, float], set[int], set[str], set[bool]]:
+    def trust_domain(self) -> tuple[float, float] | set[int] | set[str] | set[bool]:
         """
         If _trust_domain is None, then we trust the entire domain (so we return the domain).
         """
@@ -235,7 +235,7 @@ def trust_domain(self, value):
                 # The search domain must stay a subset of the trust domain, so set it as the intersection.
                 self.search_domain = intersection(self.search_domain, value)
 
-        self._trust_domain = cast(Union[tuple[float, float], set[int], set[str], set[bool]], value)
+        self._trust_domain = cast(tuple[float, float] | set[int] | set[str] | set[bool], value)
 
     @property
     def domain(self) -> tuple[float, float] | set[int] | set[str] | set[bool]:
diff --git a/src/blop/utils/sets.py b/src/blop/utils/sets.py
index a92fa77..c99e251 100644
--- a/src/blop/utils/sets.py
+++ b/src/blop/utils/sets.py
@@ -1,9 +1,9 @@
-from typing import Union, cast
+from typing import cast
 
 import numpy as np
 
 
-def validate_set(s, type="continuous") -> Union[set, tuple[float, float]]:
+def validate_set(s, type="continuous") -> set | tuple[float, float]:
     """
     Check
     """
@@ -12,7 +12,7 @@ def validate_set(s, type="continuous") -> Union[set, tuple[float, float]]:
             try:
                 x1, x2 = float(s[0]), float(s[1])
                 if x1 <= x2:
-                    return cast(Union[tuple[float, float]], (x1, x2))
+                    return cast(tuple[float, float], (x1, x2))
             except Exception:
                 pass
         raise ValueError(