diff --git a/.github/workflows/suggest-reviewers.yml b/.github/workflows/suggest-reviewers.yml
new file mode 100644
index 0000000000..fee541d09f
--- /dev/null
+++ b/.github/workflows/suggest-reviewers.yml
@@ -0,0 +1,35 @@
+name: suggest-reviewers
+
+on:
+  pull_request:
+    types: [opened, synchronize, ready_for_review]
+
+permissions:
+  contents: read
+  pull-requests: write
+
+jobs:
+  suggest-reviewers:
+    runs-on:
+      group: databricks-deco-testing-runner-group
+      labels: ubuntu-latest-deco
+    if: ${{ !github.event.pull_request.draft && !github.event.pull_request.head.repo.fork }}
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        with:
+          fetch-depth: 0
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@eac588ad8def6316056a12d4907a9d4d84ff7a3b # v7.3.0
+        with:
+          version: "0.6.5"
+
+      - name: Suggest reviewers
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          GITHUB_REPOSITORY: ${{ github.repository }}
+          PR_NUMBER: ${{ github.event.pull_request.number }}
+          PR_AUTHOR: ${{ github.event.pull_request.user.login }}
+        run: uv run tools/suggest_reviewers.py
diff --git a/tools/suggest_reviewers.py b/tools/suggest_reviewers.py
new file mode 100644
index 0000000000..cf27ba8ca8
--- /dev/null
+++ b/tools/suggest_reviewers.py
@@ -0,0 +1,297 @@
+#!/usr/bin/env python3
+# /// script
+# requires-python = ">=3.12"
+# ///
+
+import fnmatch
+import os
+import subprocess
+import sys
+from datetime import datetime, timezone
+from pathlib import Path
+
+MENTION_REVIEWERS = True
+CODEOWNERS_LINK = "[CODEOWNERS](.github/CODEOWNERS)"
+MARKER = "<!-- REVIEWER_SUGGESTION -->"
+_login_cache: dict[str, str | None] = {}
+
+
+def classify_file(path: str) -> float:
+    p = Path(path)
+    if p.name.startswith("out.") or p.name == "output.txt":
+        return 0.0
+    if path.startswith(("acceptance/", "integration/")):
+        return 0.2
+    if path.endswith("_test.go"):
+        return 0.3
+    return 1.0 if path.endswith(".go") else 0.5
+
+
+def get_changed_files(pr_number: str) -> list[str]:
+    r = subprocess.run(
+        ["gh", "pr", "diff", "--name-only", pr_number],
+        capture_output=True,
+        encoding="utf-8",
+    )
+    if r.returncode != 0:
+        print(f"gh pr diff failed: {r.stderr.strip()}", file=sys.stderr)
+        sys.exit(1)
+    return [f.strip() for f in r.stdout.splitlines() if f.strip()]
+
+
+def git_log(path: str) -> list[tuple[str, str, datetime]]:
+    r = subprocess.run(
+        ["git", "log", "-50", "--no-merges", "--since=12 months ago", "--format=%H|%an|%aI", "--", path],
+        capture_output=True,
+        encoding="utf-8",
+    )
+    if r.returncode != 0:
+        return []
+    entries = []
+    for line in r.stdout.splitlines():
+        parts = line.strip().split("|", 2)
+        if len(parts) != 3:
+            continue
+        try:
+            entries.append((parts[0], parts[1], datetime.fromisoformat(parts[2])))
+        except ValueError:
+            continue
+    return entries
+
+
+def resolve_login(repo: str, sha: str, author_name: str) -> str | None:
+    if author_name in _login_cache:
+        return _login_cache[author_name]
+    r = subprocess.run(
+        ["gh", "api", f"repos/{repo}/commits/{sha}", "--jq", ".author.login"],
+        capture_output=True,
+        encoding="utf-8",
+    )
+    login = r.stdout.strip() or None if r.returncode == 0 else None
+    _login_cache[author_name] = login
+    return login
+
+
+def _codeowners_match(pattern: str, filepath: str) -> bool:
+    if pattern.startswith("/"):
+        pattern = pattern[1:]
+        if pattern.endswith("/"):
+            return filepath.startswith(pattern)
+        return fnmatch.fnmatch(filepath, pattern) or filepath == pattern
+    return fnmatch.fnmatch(filepath, pattern) or fnmatch.fnmatch(Path(filepath).name, pattern)
+
+
+def parse_codeowners(changed_files: list[str]) -> list[str]:
+    path = Path(".github/CODEOWNERS")
+    if not path.exists():
+        return []
+    rules: list[tuple[str, list[str]]] = []
+    for line in path.read_text().splitlines():
+        line = line.strip()
+        if not line or line.startswith("#"):
+            continue
+        parts = line.split()
+        owners = [p for p in parts[1:] if p.startswith("@")]
+        if len(parts) >= 2 and owners:
+            rules.append((parts[0], owners))
+
+    all_owners: set[str] = set()
+    for filepath in changed_files:
+        matched = []
+        for pattern, owners in rules:
+            if _codeowners_match(pattern, filepath):
+                matched = owners
+        all_owners.update(matched)
+    return sorted(all_owners)
+
+
+def score_contributors(
+    files: list[str], pr_author: str, now: datetime, repo: str
+) -> tuple[dict[str, float], dict[str, dict[str, float]], int]:
+    scores: dict[str, float] = {}
+    dir_scores: dict[str, dict[str, float]] = {}
+    scored_count = 0
+    author_login = pr_author.lower()
+
+    for filepath in files:
+        weight = classify_file(filepath)
+        if weight == 0.0:
+            continue
+        history = git_log(filepath)
+        if not history:
+            parent = str(Path(filepath).parent)
+            if parent and parent != ".":
+                history = git_log(parent)
+        if not history:
+            continue
+
+        top_dir = str(Path(filepath).parent) or "."
+        file_contributed = False
+        for sha, name, commit_date in history:
+            if name.endswith("[bot]"):
+                continue
+            login = resolve_login(repo, sha, name)
+            if not login or login.lower() == author_login:
+                continue
+            days_ago = max(0, (now - commit_date).total_seconds() / 86400)
+            s = weight * (0.5 ** (days_ago / 150))
+            scores[login] = scores.get(login, 0) + s
+            dir_scores.setdefault(login, {})
+            dir_scores[login][top_dir] = dir_scores[login].get(top_dir, 0) + s
+            file_contributed = True
+        if file_contributed:
+            scored_count += 1
+    return scores, dir_scores, scored_count
+
+
+def top_dirs(ds: dict[str, float], n: int = 3) -> list[str]:
+    return [d for d, _ in sorted(ds.items(), key=lambda x: -x[1])[:n]]
+
+
+def fmt_reviewer(login: str, dirs: list[str]) -> str:
+    mention = f"@{login}" if MENTION_REVIEWERS else login
+    return f"- {mention} -- recent work in {', '.join(f'`{d}/`' for d in dirs)}"
+
+
+def select_reviewers(ss: list[tuple[str, float]]) -> list[tuple[str, float]]:
+    if not ss:
+        return []
+    out = [ss[0]]
+    if len(ss) >= 2 and ss[0][1] < 1.5 * ss[1][1]:
+        out.append(ss[1])
+        if len(ss) >= 3 and ss[1][1] < 1.5 * ss[2][1]:
+            out.append(ss[2])
+    return out
+
+
+def compute_confidence(ss: list[tuple[str, float]], scored_count: int) -> str:
+    if scored_count < 3 or len(ss) < 2:
+        return "low"
+    if len(ss) >= 3 and ss[0][1] > 2 * ss[2][1]:
+        return "high"
+    if len(ss) >= 3 and ss[0][1] > 1.5 * ss[2][1]:
+        return "medium"
+    return "low"
+
+
+def fmt_eligible(owners: list[str]) -> str:
+    if MENTION_REVIEWERS:
+        return ", ".join(owners)
+    return ", ".join(o.lstrip("@") for o in owners)
+
+
+def build_comment(
+    sorted_scores: list[tuple[str, float]],
+    dir_scores: dict[str, dict[str, float]],
+    total_files: int,
+    scored_count: int,
+    eligible_owners: list[str],
+    pr_author: str,
+) -> str:
+    reviewers = select_reviewers(sorted_scores)
+    suggested_logins = {login.lower() for login, _ in reviewers}
+    eligible = [
+        o
+        for o in eligible_owners
+        if o.lstrip("@").lower() != pr_author.lower() and o.lstrip("@").lower() not in suggested_logins
+    ]
+
+    lines = [MARKER]
+    if reviewers:
+        lines += [
+            "## Suggested reviewers",
+            "",
+            "Based on git history of the changed files, these people are best suited to review:",
+            "",
+        ]
+        for login, _ in reviewers:
+            lines.append(fmt_reviewer(login, top_dirs(dir_scores.get(login, {}))))
+        lines += ["", f"Confidence: {compute_confidence(sorted_scores, scored_count)}"]
+        if eligible:
+            lines += [
+                "",
+                "## Eligible reviewers",
+                "",
+                "Based on CODEOWNERS, these people or teams could also review:",
+                "",
+                fmt_eligible(eligible),
+            ]
+    elif eligible:
+        lines += [
+            "## Eligible reviewers",
+            "",
+            "Could not determine reviewers from git history. Based on CODEOWNERS, these people or teams could review:",
+            "",
+            fmt_eligible(eligible),
+        ]
+    else:
+        lines += [
+            "## Suggested reviewers",
+            "",
+            f"Could not determine reviewers from git history. Please pick from {CODEOWNERS_LINK}.",
+        ]
+
+    lines += [
+        "",
+        f"<sub>Suggestions based on git history of {total_files} changed files "
+        f"({scored_count} scored). "
+        f"See {CODEOWNERS_LINK} for path-specific ownership rules.</sub>",
+    ]
+    return "\n".join(lines) + "\n"
+
+
+def find_existing_comment(repo: str, pr_number: str) -> str | None:
+    r = subprocess.run(
+        [
+            "gh",
+            "api",
+            f"repos/{repo}/issues/{pr_number}/comments",
+            "--paginate",
+            "--jq",
+            f'.[] | select(.body | contains("{MARKER}")) | .id',
+        ],
+        capture_output=True,
+        encoding="utf-8",
+    )
+    if r.returncode != 0:
+        print(f"gh api comments failed: {r.stderr.strip()}", file=sys.stderr)
+        sys.exit(1)
+    for cid in r.stdout.splitlines():
+        cid = cid.strip()
+        if cid:
+            return cid
+    return None
+
+
+def main():
+    repo = os.environ["GITHUB_REPOSITORY"]
+    pr_number = os.environ["PR_NUMBER"]
+    pr_author = os.environ["PR_AUTHOR"]
+
+    files = get_changed_files(pr_number)
+    if not files:
+        print("No changed files found.")
+        return
+
+    now = datetime.now(timezone.utc)
+    scores, dir_scores, scored_count = score_contributors(files, pr_author, now, repo)
+    sorted_scores = sorted(scores.items(), key=lambda x: -x[1])
+    eligible = parse_codeowners(files)
+    comment = build_comment(sorted_scores, dir_scores, len(files), scored_count, eligible, pr_author)
+
+    print(comment)
+    existing_id = find_existing_comment(repo, pr_number)
+    if existing_id:
+        subprocess.run(
+            ["gh", "api", f"repos/{repo}/issues/comments/{existing_id}", "-X", "PATCH", "-f", f"body={comment}"],
+            check=True,
+        )
+    else:
+        subprocess.run(
+            ["gh", "pr", "comment", pr_number, "--body", comment],
+            check=True,
+        )
+
+
+if __name__ == "__main__":
+    main()